Conferência Anual do Banco Central 2026 - 13/05 (PT/EN)

[music]
[music]
economic
economist.
Central Brazil.
Gabriel Galipo, President Bank of
Brazil.
[applause]
[applause]
special.
involving
collabor
This event.
Fore
Central
organisms.
Say
French.
for
Central.
Central.
Shock.
forch.
Foreil,
braz.
Experience
moment. Fore
speech.
So cancept
form.
forch.
foreign.
Uh
[clears throat]
she
Briato. [applause]
>> [applause]
>> Brazil.
>> [applause]
>> Senior Christopher
Monar International
[applause]
[applause]
Well, good morning everybody. I'll
switch to English now. My job here is to
introduce our keynote speaker and I'll
do that uh quickly because we have a
full schedule until the end of the week
including ordinary sessions, other
speeches and this time the novelty of
poster sessions as well during the
intervals. So uh I'll just briefly
mention that professor ers got his PhD
from the University of Chicago worked at
the Fed for some time and currently he's
a deputy director for monetary and
capital markets uh department in the
international monetary fund and
throughout his career he has researched
and published extensively at a high
level on topics related to
microeconomics omics including open
macroeconomics, macro credential,
uh monetary policy and several others.
So if you heard Governor Gallipol's uh
opening remarks, you see that we are now
facing the fourth supply shock in a very
short period of time. And in this
regard, his talk couldn't be more timely
because he will address this issue. But
as always on a very uh rigorous and
analytical perspective. So I want to
thank professors for very kindly
accepting our invitation coming down
here and providing us our speech. So
professors the floor is yours. Thank you
very much.
[applause]
So I want to begin by um thanking the
central bank of Brazil uh and uh
governor uh Galipolo and uh deputy
governor Picetti for uh inviting me to
speak to you at this really fantastic uh
annual resource conference.
I am uh really deeply honored to be here
and thank uh Paulo also for the very
kind introduction. the usual disclaimer
is going to apply. So these uh views are
my own and not those of the
International Monetary Fund.
So um few people really could have
anticipated at the beginning of this
decade some of the challenges that
policymakers now confront. So this
includes a heightened geopolitical and
economic fragmentation,
but also the need to really skillfully
navigate the sort of challenges uh that
are uh associated with new technologies,
radically new technologies that will
affect how we make payments uh that will
really affect how we work and interact
together.
So you it really is um important to
think about uh research in this sort of
environment. And so in that vein, your
uh work here uh in uh the research side
really will be invaluable in helping
central banks and other policy in
institutions navigate uh these uh these
transformative changes.
So one major and consequential shift um
following up on uh Governor Golo's talk
in the macro environment is the
increasing prominence of supply shocks
and that'll be the focus of my lecture
today. So it's really striking um that
the framework reviews conducted by major
central banks at the beginning of this
decade devoted very little attention to
supply shocks. they were almost
exclusively focused on the problems
posed by very low interest rates and the
effective lower bound on policy rates.
So it's all about how to stimulate
demand and now um supply shocks are
likely to uh essentially move into a
much much more prominent role where uh
their prominence is probably the highest
that they've seen since the late uh
1970s.
So and rather than being confined to a
particular sector,
we see uh supply shocks is very
broad-based. They uh are associated with
big uh geopolitical shifts that have
affected energy markets.
A move away from open trade is
illustrated by the Baker Bloom and Davis
index.
The restructuring of global supply
chains and climate induced food shocks.
So there's really good reason to think
that we might be entering an environment
in which shocks are layered upon shocks
or what you might call a great
immodderation.
So the pro
[laughter]
so the the prospect of a shock on shock
environment um really um raises
questions both for policy transmission
for policy strategy and indeed for the
conduct of inflation targeting.
So in terms of transmission really a key
issue is um that most of the literature
is focused very much on linear models
where you're looking at the effects of
shocks in isolation
but uh the pandemic has really uh
pointed us to thinking about the
importance of nonlinearities in
interactions between shocks. So in
particular can shocks transmit
differently if inflation is running high
before the shock or output is uh running
very strong. So you have a lot of
resource pressures. So this has been a
an issue for a long time for central
banks. Um if we go back um to the
experience of the 1970s, the OPEC shocks
were preceded by big food shocks and
it's very likely that the interactions
between them amplified the effects. But
it's certainly very much uh in focus of
central banks today where they have to
think about dealing with big energy
shocks and think about whether
transmission could be affected by uh the
uh developments around the inflation
surge. So high inflation in the last few
years will that amplify the transmission
of energy shocks.
So uh an environment uh of proliferating
shocks also matters a lot for policy
strategy.
So advanced economy central banks at
least really have typically looked
through transient supply shocks and uh
really the grounds for this were very
intuitive. If they believed that the
shock was transient, then it made little
sense to react to it because you really
couldn't speed the re return of
inflation to target very much and you
just give a big hit to the real economy.
Now, conversely, uh emerging markets
took a very different tack and they
really thought it was important uh in
the context of the inflation surge to
respond very aggressively to these uh
shocks in order to uh instill a strong
credibility.
But um it really raises the question. So
what are the lessons of the 2021
inflation surge? Is it that all central
banks should essentially respond more
aggressively to inflation or do we need
a lot more nuance in that? Uh and it we
want to make sure that we avoid um the
opposite mistake of essentially causing
needless recessions.
Now related issues on the merits of a
gradualist approach. So if inflation
rises persistently above target, what's
the best way to bring it back down? Now
a gradualist approach has obvious appeal
because uh essentially uh it cushions
the blow to the real economy, but it can
make the um economy much more vulnerable
uh to uh subsequent inflation shocks.
And so you have to think about uh what
are the uh sort of uh risks associated
with leaving inflation high and only
bringing it down very slowly. So the
riskmanagement considerations are really
paramount here. So this is an old
question about how to navigate these
tradeoffs uh a gradualist approach or a
more forceful approach but it has new
relevance in our current environment
and then uh a more shockprone
environment can also have big
implications for the conduct of
inflation targeting. So I'll say at the
onset that you know I think inflation
targeting is very well suited for a more
shockprone environment. Uh the framework
has proved remarkably durable and uh in
particular uh has yielded impressive
outcomes in controlling inflation in
both advanced and emerging market
economies.
And uh this uh strong uh performance
partly reflects a strong government and
transparency act uh aspects of inflation
targeting which um have improved policy
effectiveness but also really uh
enhanced uh public uh accountability and
all this really helps shore up uh
support central bank independence but
there may be changes required at the
margin. So in that vein, a key open
question is whether uh a more shockprone
environment will um really diminish the
ability of central banks to practice
flexible inflation targeting meaning to
put a very significant weight on the uh
activity or employment uh component uh
of their of their goals. It might be a
secondary mandate for many central
banks.
In a very interesting recent paper,
Claudia Boreio uh has really traced the
history of uh fin uh of uh inflation
targeting over several decades. And one
key feature of his analysis was that you
see a progressive shift towards more and
more flexible uh inflation targeting. So
central banks as they gain credibility
were able to uh adjust uh policy to take
account of these uh goals uh for
activity and employment. But you know
that that might well change uh going
forward. Uh credibility started out uh
low in uh the 1990s. Um and it was
really critical to make sure that
central banks prioritized inflation
control to build up that credibility.
But good luck also helped. I mean so
this was an environment in which uh
really central banks benefited from much
smaller shocks and a lot of favorable
developments including essentially uh
very favorable cost shocks associated
with uh the big uh scaling up of imports
from uh emerging market economies.
Now as uh it matured, central banks
could focus a lot more heavily on the
employment component of their mandates
as Lars. Spencson has argued and um this
you know means that they could look uh
at returning inflation to target over a
somewhat longer horizon.
And the the main concern later then
became weak weak demand and how to
support weak demand as I alluded to
earlier. But moving forward, uh, central
banks really have to think about they
have a lot more credibility, but how do
they maintain that credibility in an
environment of much bigger shocks? So,
we could well see inflation deviate a
lot more from targets and more
persistently from target. It could be uh
under the sort of reaction functions
that have been followed in the last
couple decades. Something like an
inflation forecastbased rule. It can be
much more pronounced deviations of
inflation from target and longer
horizons to get back and that
potentially can undermine central bank
credibility. So do we need changes in
strategy in this sort of environment?
So moving uh to my uh kind of the more
grist of my uh presentation, these are a
lot of highle considerations that I'll
try to address. Um and I'll devote most
of my attention to using a structural
modeling framework that um is really
aimed at better understanding some of
the risks posed by supply shocks. It's
based very much on work with my
colleague at the fund Jesper Linde and
Matias Trapant of Gertie University
and you I must say at the onset that uh
I wouldn't really have been able to to
do this work without their invaluable
contributions uh both in terms of
economic insights but also uh really
their amazing quantitative skills
because this was uh a major quantitative
effort. So uh our model then is going to
try to capture some of the features that
we think are really relevant for
accounting for the inflation surge uh in
the United States and uh many other
economies. I'll focus in the calibration
of the model on the United States, but I
think it really does have uh quite
broader applicability. And this includes
misperceptions about the persistence of
shocks. So in our modeling framework as
I'll describe in more detail agents
can't tell if shocks are uh persistent
or much more transient and have to
essentially I use common filtering to
make inferences.
We'll have various forms of
nonlinearities in Philips curve. I'll
describe more but it'll include a form
of regime shift shifting where
essentially uh the uh intrinsic
persistence of the economy shifts
endogenously
and then uh after describing the model
and I'll discuss some extensions uh to
open economies that include emerging
market economies to think about uh how
um some of the uh supply shocks that
we'll see in the closed economy model
would play out in an open economy
setting.
So I think overall this framework should
be very useful to central banks thinking
about um how to um communicate about uh
the sort of uncertainty posed by supply
shocks including through alternative
scenarios in alt sims as well as uh
hopefully refine policy advice in
dealing with these sort of shocks. And
so kind of specifically I'll look at the
um implications of prescriptions such as
looking through supply shocks when does
it make sense when does it not? And
finally I'll discuss some possible
implications for inflation targeting.
So moving now uh to the um details of
our model or some features of our model
maybe not all the details then uh our
model basically builds heavily on
earlier work by Harding Lindy and
Trabons. So they have a JM JME paper in
2023
that has a a number of building block
features. So it's a new Keynesian model
a GSG model with sticky prices and
wages. So the usual suspects and that
score, but also habit persistence and
consumption. And this helps the model
account for empirically reasonable
properties in response to say monetary
shocks. You get a hump-shaped response
of output.
But a key feature is that they have a
Kimble aggregator on the preference
side. And what that means uh is that uh
relative to Dixit Stiglas uh price and
wage Phillips curve could be nonlinear.
And in particular, it's going to really
affect the transmission of uh both
demand and cost push shocks. And so this
is illustrated just by borrowing some uh
results from their paper where what we
show here on the uh vertical axis is
going to be inflation. The output gap so
resource pressure on the horizontal with
the left side meaning more resource
pressure. And this blue mash of points
right here is just derived from a
stochastic simulation where we look at
the effects of demand shocks mainly
monetary policy shocks in the linearized
version of the model. So not surprising
this is a straight line. If you have a
cost shock against that environment then
it's just going to shift up the line
relative to this median point. Uh and
it's not going to depend on resource
slack or the level of inflation. It's
just a one-off shift relative median.
Now, it's very different in uh their
environment with a nonlinear Phillips
curve. And in particular, uh they dubbed
it a banana shape because that's what it
looks like. It's uh levels of uh well,
when the economy is in recession, the
output gap is negative. You have a very
flat Phillips curve. But when uh
resource pressures are much higher, then
you can see the Phillips curve really
shifting up quite a bit right here. So
really much more upward slope in a hot
economy. And that also has very
important implications for cost shocks.
And in particular, if you have this
one-off cost shock, it shifts up the
curve just a little bit uh at uh very
low levels of resource pressure when you
have a lot of slack in the economy, but
much much more. So you have a lot of
pass through in a an economy that is
much hotter. And this reflects uh
strategic complimentarities and price
setting uh associated with this uh
kimble aggregate.
Moreover, we get very similar sort of
effects with uh uh cost push shots uh
depending on the level of inflation that
basically the pass through a lot higher
in that case. Oh yeah. Oh sure. Just uh
if I move away from the Yeah. Oh no
problem at all. Thanks.
So in terms of what we do in uh the way
of uh extra value added our bottle is
going to introduce three new features.
First we have Coleman uh learning about
the persistence of the shocks and uh
here uh as I mentioned we have uh a
framework where agents can't tell
whether shocks are going to be uh
persistent or transitory and uh they've
got to do uh common filtering to make
that assessment and the motivation could
be captured by uh the inflation surge
experience here uh we we focus on the US
but it was broadly applicable for many
other economies.
The outer line right here is going to be
USPCE inflation uh over the inflation
surge period. What the purple segments
are showing would be the forecast of uh
the survey of professional forecast. So
it's a median forecast. It's quarterly
at an annualized rate and you can see
these nose diving back to 2% uh even
though inflation's going to the 5%
range. So forecasters really thought
that this is going to be a transient
shock. And um this is really a key
aspect that we're really trying to
capture in our model is to uh account
for this uh fact that uh inflation went
up a lot but uh forecasters really
thought it was going to be transient for
quite some time. We also uh allow for an
inflation forecastbased tailor rule with
smoothing and this plays a very
important role in our analysis. Often
times the man monetary rule is almost an
afterthought but here it's very uh
critical. So uh inflation is going to
respond to the uh expected path of
inflation four quarters ahead. And this
is consistent with really a voluminous
literature on forward-looking tailor
rules. It really makes sense to think of
inflation targeting as looking at a
forecast of inflation. And so a tailor
rule makes quite a bit of sense in that
context. The implications though are
also very much aligned with uh policy
during the inflation surge where you can
see uh looking at the three-month
forecast of uh the SPF participants that
the the Fed was really expected to
tighten minimally uh in early 2022 and
really even in uh well into 22 22
there's a pretty flat path for the
policy rate. They ended up raising uh
the policy rate to over 5%.
So that plays an important role. And
finally, we're going to allow for what I
call endogenous uh price and wage
indexation. This is the second really
key nonlinearity in the model. And so
you can think of the lag uh coefficient
in the Philips curve as capturing the
degree of in uh intrinsic persistence in
the economy. And what we're going to do
is make that lag coefficient instead of
being exogenous as in in many
frameworks. Then uh it's going to be
endogenous. is going to depend on the uh
level of inflation and how much it
deviates for b from baseline as well as
how persistently that occurs. And so if
you have a very persistent deviation of
inflation from baseline then uh
intrinsic resistance is going to uh
increase. So what does that mean? It
means that even small and transient
shocks uh can have pretty uh persistent
effects on inflation. Uh and of course
more persistent shocks can have really
big effects on inflation as we'll see.
So getting into um some of the uh
details of the model
the period utility functions can be
assumed to be logarithmic consumption of
individual agents is CT we have habit so
that's why external consumption uh
enters with uh a capital CT and then
there's disutility from uh working and
that's NT the household budget
constraint is just very familiar
households consume uh the nominal
spending is PTC CT. They buy nominal
bonds. They make labor income. They earn
uh income on their bond holdings. They
pay lumpsum taxes. They get profits. And
finally, this HAT term is just a uh
insurance uh payment or um uh sub uh an
insurance uh payment or or subsidy that
allows uh consumption to be equalized
across agents because you have a
staggered wage setting. There's a
unequal distribution of income. This
means everybody gets the same
consumption.
Now turning to firms, uh the setup is
really familiar in most respects. So we
have perfectly competitive firms. They
bundle intermediate goods into a final
good according to a Kimble aggregator.
And so this is going to give rise to
demand functions for the uh intermediate
good where the demand for any individual
producer I relative to the aggregate is
going to depend inversely on the price
it charges relative to the aggregate.
And then you have this CYP term and that
CYP term is going to give you the kink
in the Philips curve, the nonlinearity
that I described earlier.
Otherwise, the problem is is really
quite standard. Monopolistically
competitors produce monopolistically
competitive intermediate producers are
going to maximize intermporal profits uh
subject to a cowboy style friction as
long as they get the uh ability to uh
reoptimize their price. And those that
don't, I'll describe their problem in a
minute. that's going to give rise to the
indogexation feature.
Now I mentioned that uh agents have to
learn about the underlying shocks and so
you know here we have a cost push shock
a that's going to consist of an iid part
a and a persistent part a and agents can
observe the sum but not the separate
components.
So um this basically gives rise to a
simple state space representation where
uh agents are going to use the cop
column filter to do optimal forecasting
and that's how they predict both at and
ape
is going to be that the uh innovation
variance of the transitory component is
a lot bigger than that of the persistent
component. What does that mean? It means
that uh when they see a persistent shot
they're going to think it's transitory.
So now the monetary policy rule as I
mentioned earlier it's a forecast based
tailor rule where uh the policy rate is
responding to the inflation forecast uh
expectation of pit t plus4 uh the output
gap and then there's smoothing and we do
consider variations where the central
bank responds to inflation or to uh
inflation eight quarters ahead.
So now we move to this uh endogenous
price and wage indexation mechanism
and um really uh it's starts out kind of
familiar because what's going to happen
is that those firms that can't optimize
set their price as a markup pit t
squiggle over the price they charge last
period where pi t squiggle is going to
be a weighted average of um uh the
inflation target pi and of last period's
inflation pi t minus one with a weight
kai t in the latter. So that's just
familiar from Cristiano Wikbomb Evans.
The difference here is that kite t is
going to depend on a geometric weighted
average pi t star uh of lagged inflation
pi t p I'm sorry pi t minus one. So it's
going to be a long distributed average
of past inflation. So this is going to
capture the idea that inflation goes up
very persistently then that indexation
parameter is going to shift and it
really changes the structure of the
economy. It really very marketkedly
affects the Philips curve. So you can
see that uh visually by uh our our
calibrated uh indexation function right
here where we show the level of
indexation. You can't really read that
on the left uh axis and on the vertical
axis it's going to show uh geometric
average of inflation. So if inflation's
from 0 to 2%, the indexation parameter
is set equal to zero. If inflation's at
three, it's still extremely low. But 0 2
it essentially doesn't matter. Um but if
inflation gets to four or 5%, then it
really starts to affect the dynamics. Uh
inflation indexation goes to 7 or 08. So
uh that means that these transient
shocks can have much more persistent
effects.
Now I want to uh caution that indexation
shouldn't be viewed as limited to formal
indexation schemes. We're really
interpreting it much more broadly as how
inflation dynamics can shift in a higher
inflation environment. And so I think
it's very much in the spirit of uh some
other work including for instance a very
nice paper by Claudia Boreo and
colleagues at BIS uh on two inflation
regimes where they distinguish between
inflation behavior in a low inflation
environment where uh sector specific
shocks have very little influence on
other sectors of the economy where
transmission of shocks to wages uh is is
very low. Um but uh in a hot uh e
economy uh where inflation is high that
transmission changes really quite
marketkedly and in particular sector
specific shocks can have big spillovers
to other sectors of the economy and uh
shocks that affect uh uh inflation can
spill over a lot to wages. So workers
become much more resistant to uh wage
declines when they've already taken big
hits. And so this is you really what
we're aiming to capture here.
So to get a better sense of what it
implies, uh what we show here is a
geometric average of US past inflation
going back to the early 60s. And my
apology this 0.02 is quarterly
inflation. It has to be annualized. So
we forgot to do that. Uh so this is like
uh essentially 8 to 10% inflation here.
That geometric average is very high in
the 70s and uh until the early 80s and
then comes down a lot. And so those
features are inherited by the indexation
parameter which is very low through the
60s uh till till basically the late 60s
and then it starts to jump as inflation
is high and then indexation does come
down a lot uh first in the 80s and then
precipitously in the 90s and basically
it stays at very low levels uh for the
next 30 years and this is consistent
with a lot of evidence through the great
moderation period of very low
indexation. So that was almost an
invariable uh property of empirical
estimates very low indexation but you
can see that this is particular to this
environment and in the more recent
period we've seen a big jump in
indexation it's come down uh
significantly recently.
So I won't spend much time on the
calibration just uh aside from saying
that we have a very conventional uh
calibration of the tailor rule. So uh
the uh smoothing coefficient is 085. The
coefficient on expected inflation is one
and a half and uh it's 0.5 on the output
gap when annualized.
The other aspect as I mentioned is that
the standard deviation of the transitory
markup shock is a lot bigger than that
of the persistent shock and we'll see
what that implies in just a moment.
So um now moving to uh the model. So in
our paper um what we have is
and essentially I try to account for the
inflation surge experience. And so we
feed in shocks both to supply and demand
that uh try to account as as well as
possible for the inflation surge. not
just the behavior of uh inflation and uh
activity in the policy rate but also uh
we really look at uh trying to fit those
expectations that I showed you earlier
but many of the features can really be
well summarized by just a one-off cost
push shock that is persistent so I'll
focus on that here in my presentation so
uh what we'll focus on is this cost push
shock with a persistence of uh 0.9 it's
an AR1 and agents are going to initially
perceive it as the transient shock.
So what we see looking at the left
panel, this shows inflation. The outer
envelope, the red lines show the path of
actual inflation implied by our model.
And so it's very similar to the sort of
hump-shaped response that I showed you
uh in reality. Then you see these green
lines right here, and what they are are
the uh forecasts of agents within the
model. And so very similar to the data
that we showed you earlier at each date
they forecast that inflation's going to
go back quite quickly and it's only uh
when they are continually surprised that
they're updating their common filter and
say hey maybe this is a more persistent
shock. Now in terms of the policy rate
response there's very little action in
the policy rate in the first year. Why?
because the central bank uh also
believes that the talk is going to be
transient and so the net effect is that
uh you actually see that uh output
expands. Why is that? Because in the
very near term the real interest rate
falls. This is a pretty passive uh
policy rate response. The real rate
falls, output actually expands. So you
get uh a cost push shock uh actually uh
not only boosting inflation but actually
uh boosting output a little bit. And you
I think that this you know can help
account for uh the resilience of many
economies that we observed during the
pandemic experience. Of course demand
shocks were also important. Um but uh
moreover it's also an interesting thing
to contemplate when identifying
structural bars because a key uh
identifying uh restriction and sign uh
restriction methods to identify
structural bars would be to assume that
a cost push moves inflation and output
in opposite directions. And so this will
be an illustration of when that might
not uh be applicable. Uh but a really
key thing here is this humpshaped
response. I'll spend a little bit more
time on that. And it's really driven by
a combination of uh inflation going up
and that means more indexation and then
uh the the relatively passive response
of policy. Uh it's not an aggressive
policy response and uh that helps uh
fuel this hubshape boom.
So uh misperceptions then can have a lot
worse consequences in a nonlinear
modeling framework. And um we'll we'll
start to look at the implications by
start by looking at the implications in
a linearized model of the same shock. So
the same cost push shock with the same
common learning is going to give rise to
this inflation response uh where
inflation just goes up a little bit
above uh 3% shown by the uh dashed
purple line here
that's pretty easy to control. It
doesn't pose a big challenge for
monetary policy. Now if you allow for
the um
Kimble aggregator specification so you
have that form of nonlinearity then you
go to this red dash line right here and
you see that the response is amplified
quite a bit but it really doesn't
increase the persistence a lot. It comes
back down pretty fast. It's only when
you have these two forms of nonlinearity
that you really get this big humpshaped
response and this poses a much bigger
policy challenge. So um in in terms of
thinking about the policy challenges
here, essentially what the policy maker
can do is they're facing a much more
vertical Phillips curve. So inflation
goes up a lot faster if the economy is
hot or if you have big cost push shots,
but it's pretty easy to bring it down.
You have a big kind of kappa in your
your uh Phillips curve equation, a big
slope of the Philips curve. And so you
you can just contract output a little
bit. Inflation comes right back down. So
you just kind of ride back down a
vertical Phillips curve. In contrast, in
this case, once inflation's high right
here, then uh the you might have some
latitude uh because the economy is
overheating to uh bring down inflation.
But still, even once you brought output
to potential, you've got a big inflation
problem. And so it really is going to
mean that the tradeoffs are are much
much worse in this environment, as we'll
see in a moment.
Now um the nonlinearities are really
going to matter under two conditions.
One, if you're at the baseline, the
shocks are big. Uh the second is you
start from poor initial conditions. So
inflation's already above uh targets and
then you have interactive effects and
then the shocks wouldn't have to be as
big. And so we illustrate this in the
first uh case where you start from the
baseline right right here where I just
repeat the last uh diagram where
inflation goes up to almost 7% here in
the linearized model about 3%. So that's
with uh the same shock as I showed you a
second ago. But if the shock is half as
large then the nonlinearities aren't
nearly as big. And so that's shown in
the lower left panel where you see uh in
the model with both nonlinearities
inflation only goes up uh to a little
over 3%. And uh in the linearized model
there's not much of a gap. So given that
uh US inflation really didn't average
over a 30 some year period above 3% then
you might think that uh 9 million years
weren't particularly consequential
during a lot of that period. And so
during the great moderation period, it
made sense to think about uh very little
indexation and using forecast-based
rules and that these would be effective.
But the question is, you know, what
happens when you when you move out of
that period when if the the shocks
really are a lot bigger.
Weak initial conditions, you know, can
work very similarly. So what we do here
is consider
um allowing for um shocks that boost
output well above potential as well as
boost inflation above target. So and
then consider a cost push shock on top
of that. So if we start from a baseline
the steady state the cost push shock is
going to push uh inflation 4.7% above
its target level. So that's how the the
shock is scaled here. But if you look at
the [clears throat] uh fourth line for
instance uh the output gap is 4.1%
inflation uh 2.7 again induced by these
uh prior shocks and then the inflation
response for the same cost per shock
goes to over 8%. So you almost double
the size of the shock against this
backdrop. So you this really is
underscoring how nonlinearities can can
matter quite a bit through these
interactive effects.
So, you know, really a key issue um is
trying to assess with more precision
when nonlinearities kick in and really
materially materially alter
transmission. So, is it when uh you have
3% inflation, you get hit by energy
stock that it matters or do you have to
go to four or 5% inflation for that to
happen? And you know this really is
going to depend a lot in our model on
the uh calibration or estimation of the
indexation function. And so uh what
we've uh been working on is in uh is on
estimating our model uh to try to uh pin
down this indexation function to a
higher degree. And uh one of course big
challenge is that there have been very
few episodes of high inflation in US
history outside of the inflation surge
unless you go back to the uh period of
the 1960s and 1970s. So that's
essentially exactly what we've done is
go way back to the early 60s to try to
come up with more refined estimates of
the indexation function as well as the
underlying shocks. So at this point
we've partially estimated the model um
meaning that we've uh fit um the uh
model to data on activity uh so output
on inflation and the policy rate and
we've used a number of uh structural
shocks to do that. So we have the cost
push shock in the model. We have a
demand shock which is just a discount
factor shock in this uh simple
specification. And finally shocks to the
policy reaction function. And uh we've
uh matched uh the uh well we've
estimated the uh properties of the
underlying shock processes as well as
backed out estimates of uh each of these
components. And so that's what I'll
illustrate next.
And I think you know this is still work
in progress. it's uh very much hot off
the press. A couple of uh features
really appear to be coming uh across in
a very robust way. And one is that uh
both types of nonlinearities seem to to
matter uh during uh periods of of large
uh uh cost shocks. So that you know the
two nonlinearities I mentioned are both
very consequential and you need the
indication in addition to the gamble
aggregator. Uh and then there's a point
about nonlinearities mattering overall.
And so what what I show here is that uh
we we've got the nonlinear model with
all the shocks shown by the solid line
right here. The same shocks in the
linearized model trace out this uh
specification here. And uh the gap
between them is going to tell you about
the role of nonlinearities uh in
explaining the data. And so you see a
big role for nonlinearity starting to
kick in in the early 1970s even before
OPEC you had big food shocks um and then
you know even stronger through uh the
late '7s again in the more recent period
you see a pretty significant role for
nonlinearities.
[snorts] So uh that's uh nonlinearities
but then the period between 1990 and
2020 very little in the way of
nonlinearities. The one exception in
that period is right here in 2009 where
the linear model is going to predict a
big drop in inflation. Uh so if you have
a linear Philips curve big drop uh the
nonlinear model because of that banana
shape is going to assuage that drop and
so it actually fits the data a whole lot
better.
A second uh feature that really comes
through is in the decompositions of uh
the different uh responses including of
uh PC implation is shown here. And so
what we're going to do is uh decompose
the shocks into three components. A cost
push shock, a discount factor shock, and
a monetary policy shock. And you can see
that this bluish uh line right here
suggests that uh cost push shocks played
a really big role in the inflation runup
in the uh 70s and also in the more
recent inflation surge period. Now,
interestingly, uh negative cost shocks
actually played an important role in
restraining inflation in the great
moderation period. So you see that here.
So overall the the message is that shock
uh shocks are very consequential. Uh
there's also uh some suggestion of
important interactions between shocks
and you can see that for instance in the
recent period where uh stimulative
monetary shocks interacted with the cost
push shocks in driving up inflation.
So um I'll now uh discuss some
implications for policy strategy. So
I'll first simply assert uh result we
have the results in the paper that um
inflation forecast targeting uh is
applied to transient shocks when you're
near the steady state works really quite
well. Uh so um it doesn't make sense to
react aggressively to a a transient
supply shock when inflation expectations
are well anchored and the shock is in
fact transient. Um but you know what
really uh gives rise to the problems
then as I mentioned is when you have
more persistent shocks and so this right
here is going to repeat the uh figure uh
shown previously where inflation goes to
like 7% and output expands. And uh in
this sort of environment, it can make a
lot of sense to respond more
aggressively. And so if you think of a
rule that would put uh significant
weight on responding to realized
inflation, that's shown by the pink line
here. And it really reduces that hump
shape in inflation. There's practically
no hump. Now, there is some modest cost
to to output that's going to depend on
how quickly you shift gears and react,
but here it's really quite modest. And
if you fact in fact do a uh a simple
calculation approximating welfare using
a simple quadratic loss function with
equal weights on inflation and the
output gap. There are major benefits to
reacting more aggressively uh along the
lines of the pink line in that
environment.
So this logic is going to come out even
more forcefully
uh in the context of stoastic sims. And
uh here what we do is consider uh how um
central banks uh facing a hot economy
where inflation's gone up almost to
double digits here. Uh what what sort of
challenges they face if there are
additional shocks on top of that. So you
can think of this as really highlighting
upside inflation risks on top of an
alternative scenario. And it's probably
a little hard to see here, but what
we've done is uh basically match this uh
distribution of shocks over the the same
period that I described earlier starting
in the 60s. And you see this really huge
upward skew in the distribution of
inflation. Um and uh you know it shows
how a gradualist approach would bring
down inflation. Uh but it leaves the
economy very vulnerable. Now, of course,
a benefit of gradualism is that there's
very little hits to output. Uh, it looks
like a worse hit than actually is the
case because what's happening here is
basically the demand shocks that drive
up the inflation initially are wearing
off, but there's very little hits to
output. So, the the benefits of
gradualism would, you know, seem to be
well, we don't uh hit output a lot, but
it leaves you very very vulnerable to
these uh uh additional shocks. Now, uh
you could uh consider a more aggressive
pulse reaction function that really
brings down inflation much more quickly
as you'd expect as well as narrowing the
distribution of inflation outcomes. Now,
if you do it belatedly as here, because
you're not doing it until inflation's at
a peak, then it really could take a big
hit to output. Uh so the overall message
here is that um it you really can be uh
desirable to use a forecastbased rule
when you're near the steady state, but a
more aggressive response against uh
inflation can be desirable in response
to large shocks or when uh the economy
is uh overheating.
So we've uh also worked to basically
incorporate these same mechanisms into
an open economy uh model. We call it the
uh quantitative integrated policy
framework model. Uh it's uh been
developed by our uh MCM department. And
uh it's an open economy DSG model that
has been applied to many um both
advanced and emerging market economies.
And the model is really useful in
assessing the effects of a wide range of
shocks to energy uh prices to trade
policy to the currency risk premium. And
uh many of the sort of implications that
I've uh traced out both in terms of
transmission and policy uh implications
uh continue to apply in this uh sort of
setting. And so for example uh we might
think that currency risk premium shocks
could become a lot more uh volatile in
an environment of greater policy
uncertainty and we've shown that in uh
the recent IMF or actually pretty recent
IM IMF uh GFSR in last fall where if you
have greater economic policy uncertainty
or the VIX is higher it tends to uh
amplify exchange rate volatility.
Now these sort of shocks can have pretty
modest effects near the steady state. So
a shock to the UIP premium that uh
drives up uh uh the or sorry depreciates
the currency by 10%. But have pretty
modest effects on inflation uh including
in small open economies and in in in uh
open uh emerging market economies in
normal times and be moderately
stimulative for output. Uh but that can
change a lot against uh the backdrop of
an economy where inflation is already
well above target. And in fact there's a
nice uh paper by Ricksbank staff
illustrating uh in Sweden's case how you
have this very state dependent
transmission occurrence at risk shocks.
Uh but this uh framework I think is very
useful for uh really developing
alternative scenarios that aren't just
focused on uh shock uncertainty. So what
happens if the currency risk premium
moves the exchange rate by 10% or oil
prices go up by 50%. But rather on
transmission uncertainty. And so that's
what I'll illustrate next uh in the
context of an energy shock.
So here we're going to think about state
trans state dependent transmission of
energy shocks using the QIPF model. And
this is basically calibrated in a very
similar way to what I described earlier
for the US economy but we've considered
many extensions uh including to small
open economies and emerging markets. So
uh here we have a 50% energy shock under
different initial conditions and uh if
you're starting out near the steady
state so inflation's near target in US
case of 2% then the 50% energy shock
moves up inflation by something like
04.5%
on average during the first year and a
little bit more thereafter.
But if you're against the backdrop of uh
an economy where the inflation rate is
running closer to 5% say four and a half
or 5% then you see really considerably
amplified transmission where the effects
on core are closer to 7. So they're not
quite double but they're getting in that
vicinity. So uh much bigger uh effects
on on core. Now, uh if you think about
the case of an emerging market, uh where
uh FX markets are are not very deep,
then uh if they're an oil importer, you
get kind of a double whammy from a
dreiation in the exchange rate when
energy prices go up. Uh and if the
foreign exchange market is not very
deep, that can have a lot of
transmission uh uh to uh the exchange
rate and in turn to inflation. Uh so,
you know, there you could even get
bigger effects. But really the key point
here is to think about uh when do these
uh nonlinearities really become
material. And so you know that's why
what we're trying to do is develop a
framework where we can think a lot more
about transmission uncertainty and bound
that uncertainty in a reasonable way
because you know we don't want to just
uh make assumptions about uh amplified
transmission that aren't rooted in data.
So the estimation of the indexation
function and other aspects of the model
should help uh help balance this sort of
uncertainty.
So uh now um I'll turn to thinking kind
of more generally about implications for
inflation targeting. And uh I think uh
in that context it's very helpful to
think about uh the uh Taylor curve. So
John Taylor made some uh really great
contributions. Uh one of which was uh
the Taylor curve. And uh so what uh just
as a refresher we have here is thinking
about how the uh unconditional
opportunity set of policy makers is
going to shift uh in a more volatile
shock environment. And this is just a
purely uh illustrative sort of uh
calibration here. But you can just
imagine that you have this should
actually be the standard deviation of
inflation on the vertical axis. the
standard deviation the output gap on um
the horizontal axis and if you think of
a point a right here this would be
associated with a particular calibration
of the tailor rule say a particular
coefficient on inflation uh or uh if
you're thinking about uh a full
commitment optimal policy a lambda
waiting uh inflation versus the output
gap that would put a relatively high
weight on the the output gap here and as
you move down along this uh curve uh
what's going to happen is
well I'm sorry I should have said if the
policy maker reacts more to inflation
then you're going to move down along
this curve so the volatility of
inflation is going to be compressed and
the volatility of output gap is going to
expand. [clears throat] Now uh if you
have a shift because of more volatile
supply shocks that's going to shift your
curve out to here so move out there. If
you keep the same reaction function,
then you're going to move up to a point
like B where you have considerably more
volatile inflation. Uh and that you
poses a big challenge for uh C uh
central banks that are inflation
targeters because it means that uh at
average inflation is going to be further
away from targets. Uh so you can imagine
that uh it can make sense to essentially
respond more forcefully and move back to
this sort of point. But it takes a shift
in the reaction function.
Now, um more difficult tradeoffs can
arise
[snorts]
if uh you're in a situation where the
policymakers opportunity set depends on
the policymakers choices. And so that's
what we're actually considering here.
And so I'm very grateful to my uh
colleague Marshian Colossa who's done a
lot of work on the QIPF model to help uh
set up these uh simulations. And what
we're trying to do is characterize
essentially conditional uh tailor
curves. So uh these are tailor curves
that would be conditional on uh
underlying shocks. And so particular we
start with a standard tailor curve in
the QIPF model derived in a very similar
way. So you could take a point A uh as
your initial point where uh the policy
maker puts some weight say one and a
half on inflation uh and then bigger
weights would uh yield uh inflation and
output gap outcomes along here and then
if you consider a bad realization of
shocks now uh what can happen is if you
stuck with the same rule sorry I did not
move that appropriately sorry uh if if
you consider the rule right here with
this bad realization of shocks then it
could be uh up at at this point with
very poor trade-offs where inflation is
very persistently high and it'll be
consistent with the specification that I
showed you earlier showing the impulse
responses but I think it's very helpful
to think of it in terms of how it
affects the opportunity set now uh if
you then uh decide to react more uh
you've got really poor trade-offs uh
it's not just moving back down the uh
nonlinear Phillips curve it you really
takes a very big hit to output to uh try
to get you back. So um the uh underlying
message is that it really may be helpful
in a more shockprone environment to uh
react more forcefully especially to
bigger shocks or in circumstances where
inflation uh has been running high uh
and you have these uh big interaction
effects.
So uh with that I'll wrap up. Um so in
my lecture I've focused on the
implications of a more shockprone
environment in which supply shocks are
larger and more pervasive than uh before
the pandemic. So I've proposed uh really
a number of tractable features on the
structural modeling side uh to help
capture how these developments may
affect transmission.
And uh I paid really particular
attention to Philips curve
nonlinearities and how they may uh
create interactions with uh shocks that
can pose major risks to price stability.
My hope is that modifications along
these lines will be very useful in
helping uh central banks uh enhance uh
their analysis of risk to the outlook
including through uh stochastic
scenarios and alternative um I'm sorry
through through alternative scenarios
and stochastic sims. But rather than
focus mainly on risk to exogenous
shocks, I really emphasize the
importance of thinking more about a
trans transmission uncertainty and
alternative scenarios and how to bound
that uncertainty, how is affected by
monetary policy choices. So this is very
relevant today as central banks must
decide how to respond to energy shocks.
Um and it'll also be important going
forward if we really are facing uh a
great moderation environment. Of course,
more research is needed to better
understand the sort of thresholds at
which these uh nonlinearities become
material. We haven't had that much
experience with it in advanced
economies, but it is important to draw
on evidence from earlier periods as well
as cross-country evidence. So on the
policy side, my uh analysis uh really
suggests more caution in looking through
supply shocks under some conditions. Um
so uh when inflation is higher the
economy is hot but at the same time it
may make a lot of sense to essentially
do so uh when you're a lot closer to
steady state and shocks really are
moderate in size and with high
probability uh likely to be transient.
So from a broader perspective, the more
volatile supply shock environment may
require some changes in the
implementation of inflation targeting uh
likely calling for a more aggressive
reaction function uh uh to inflation
under certain circumstances to maintain
policy credibility. Of course, better
characterizing when uh the risks become
sizable and how they depend uh on
country circumstances remains, you know,
really a key task for ongoing research
and was one of the reasons why uh the
research is is needed to support good
policym and I think, you know, it really
underscores uh the value of the work
here uh where um you really look forward
to the next several days of
presentations where we'll explore these
issues uh in in much greater detail.
Well, thanks very much.
[applause]
Is it on?
>> No.
>> Can I borrow your mic?
>> Oh, please.
>> Thanks a lot, Christopher. and you kept
your end of the bargain with the time,
but we uh started a little late. So, in
the benefit of time, I'm sure everybody
must have lots of questions, but we can
uh perhaps take a batch of free
questions and then you'll be around for
the rest of the week. So, let's take
three questions and uh go from that.
Fernandanda, please.
microphone
forward.
>> Thank you, Christopher. That was a very
interesting model and I think it's going
to be very useful to a number of central
banks going forward. Um but my question
is if you were a central bank and if you
agree with the scenario of um more
frequent shocks and you are not certain
on the extension of shocks like the one
you have embedded in your model
on top of the policy strategy things you
kind of penciled out here. Um should
maybe you add another one that in uh in
when facing the uncertainty on the
extent and size of the shock should
central banks always react in a more
aggressive manner.
>> Should we take a couple
questions.
Thanks for the the presentation, Chris.
Um the way I think about your sort of
endogenous persistence is maybe um as a
as a shortcut for expectation regimes
that shift between anchored and unanch
anchored right so so you have if I
understand correctly rational
expectations with the with the filtering
problem but maybe that's a shortcut for
trying to think about you know times
when expectations are anchor and
reanchor have you guys thought about
that and how much weight do you put
in the actual you know survey
expectation that we observe as as maybe
a way to detect this unanchoring um
regimes. Thank you.
I'll make a very quick quick one. Uh
it's not an easy answer. You can answer
with yes or no and then we can talk
later. But you mentioned luck is part of
what happened including the US. Um, and
do you see in your model a space for
negative cost push shock coming from AI
related productivity gains
in the horizon of monetary policy?
>> Oh, thanks for uh very good questions. I
think uh in terms of whether you'd
always want to react more aggressively
to uh a cost push shock um you I think
that uh the answer would be no. I think
if you're relatively close to the steady
state, uh inflation's uh at your target,
uh the economy is not running hot, then
uh a uh shock, whether it's an energy
shock or other form of supply shock, uh
even if it might be quite a bit more
persistent, uh you you essentially try
to calibrate bounds for how big the
effects can be. So using the energy
shock example here, you might say, well,
if it's a transient shock, then the
effects on core inflation, the 50% shock
instead of being point4 might only be
0.2. Uh but if it's uh persistent and uh
uh dies out with a halflife of say 9
months uh uh or thereabouts, then uh the
core inflation response might be more
like point4. And that's actually the
assumption underlying this. Um, but you
put reasonable bounds on it. And so it
wouldn't necessarily mean that you have
to react a lot more aggressively. I
think that you uh especially initially
you could take some time to learn about
uh really how uh big and pervasive the
shock is going to be before you uh you
know really react to it including
through uh your communication.
So um so that's one point. I think uh in
terms of the regime's wishing I think
that that is uh a uh correct
interpretation of what we have in mind
that we really have uh it uh the heart
of our model. uh the goal of trying to
capture how in higher inflation uh
environments you really have fundamental
uh changes in uh expectations about uh
how shocks will play through. It is a
rational expectations model. Uh you of
course and so what that means is that as
agents see uh indogenous persistence
increasing they're going to recognize
that shocks uh even transient shocks are
going to have a lot bigger effects. So
they internalize that and so you'd see
much more sensitivity of uh expectations
to those shocks and so it captures all
those features albeit you admit in in a
somewhat reduced form way. I think I'll
use that uh as a uh platform for saying
there are additional sorts of uh
uncertainty that we would probably want
to incorporate in other analysis. So
here we've got uh that uh a framework
where the monetary policy reaction
function is completely understood. Uh
there's uh you really no uncertainty
about the long run target of the central
bank. But you can imagine in
circumstances where inflation is running
high uh the outcomes could look even
worse than what we've shown here if you
didn't have such long run anchoring.
because it could be the case for
instance that uh if you allow inflation
to run persistently high and then react
more passively than markets expect to
that markets are going to interpret that
as a lack of commitment to your uh
long-term inflation target and so
long-term inflation expectations can
rise in that environment. So I I wrote a
paper with Andy Leaven uh many years ago
on on exactly this and it could easily
be embedded into this sort of framework.
For instance, uh we also don't have
features uh that you know I think would
be uh relevant for thinking about risks
of fiscal dominance and uh risks of
financial dominance. Um Frank Schmmetz
had a very nice talk last year in which
he discussed this in much more detail
the the former those uh and so finally
to your question about uh the
possibility for AI to exert uh negative
effects on uh inflation you know
certainly it's quite possible that we
see that and uh if we go back to the US
experience in the 2003 to 2006 period
then supply shocks really did exert a
significant negative effect on the
economy.
Um it was an environment where growth
accelerated enormously. It went from
like 2 and a.5% to 4 and a.5%. Um but
inflation was very tepid and so the the
Fed was really struggling with with very
low inflation in the uh early part of
that period and it restrained their uh
willingness to to raise interest rates.
But it it dep it really depends a lot on
the relative strength of supply versus
demand effects. their supply dominated,
the demand effects were not very big. Uh
in the uh period uh in the late 90s,
conversely, the demand effects were a
lot bigger and so the the downward
pressure on inflation, not very much.
And so if you think of the current
environment, it really is to me an open
question which uh wins. Uh so you have
demand side effects coming from uh
increases in equity valuations that
spill around all over the world in
declines in risk premium. you have
demand side effects that are positive
because you have uh data centers are
being built all over uh and you know
that creates demand pressures and so
those could be enough to offset the
deflationary impact you I think it's
it's an open question to me
>> but very good one [laughter]
>> well thank you very much Christopher we
did start our conference on a very high
note and thank you for that and I'll ask
everybody to join me in a last warm
round of applause for our guests
>> [applause]
>> Thank you very much, Mr. Christopher.
D. So,
>> [music]
[music]
>> Hey, hey, hey. [music]
[music]
[music]
[music]
>> [music]
[music]
[music]
[music]
[music]
[music]
>> Heat. Heat.
[music]
Hey. Hey. Hey.
[music]
[music]
[music]
>> [music]
[music]
[music]
[music]
[music]
[music]
>> Heat. Heat.
[music]
>> [music]
[music]
[music]
[music]
>> Heat. Heat.
[music]
>> [music]
[music]
>> Heat.
[music]
Heat.
[music]
>> [music]
[music]
[music]
[music]
[music]
[music]
[music]
[music]
>> Heat. Heat.
[music]
>> [music]
[music]
[music]
[music]
>> Heat. Heat.
[music]
[music]
[music]
>> [music]
[music]
[music]
>> Hey,
hey, hey. [music]
Heat. Heat.
[music]
[music]
[music]
>> [music]
[music]
[music]
[music]
[music]
[music]
>> Hey, hey, hey. [music]
[music]
[music]
[music]
Heat. Heat.
[music]
[music]
>> [music]
[music]
[music]
>> Heat.
Heat.
[music]
>> [music]
[music]
[music]
[music]
[music]
[music]
>> feel. Hey, hey, hey.
[music]
[music]
[music]
[music]
>> [music]
[music]
[music]
[music]
[music]
[music]
>> Hey.
Hey. Hey.
[music]
>> [music]
[music]
[music]
[music]
[music]
[music]
>> Heat. Heat. N.
[music]
[music]
Heat.
[music]
[music]
[music]
Heat.
>> [music]
[music]
[music]
[music]
>> Hey.
[music]
[music]
Hey. Hey.
Heat. Heat.
[music]
[music]
[music]
>> [music]
[music]
[music]
[music]
[music]
[music]
>> Hey, hey, hey.
>> [music]
[music]
[music]
[music]
>> Heat. Heat.
[music]
[music]
[music]
>> [music]
[music]
[music]
[music]
[music]
[music]
[music]
[music]
[music]
[music]
[music]
[music]
>> Hey. [music]
[music]
[music]
Hey. Hey.
>> [music]
[music]
[music]
[music]
>> Heat. Heat.
[music]
[music]
[music]
>> [music]
[music]
[music]
[music]
[music]
[music]
[music]
>> Hey, hey, hey.
[music]
[music]
[music]
[music]
>> [music]
[music]
[music]
[music]
[music]
[music]
[music]
>> Hey,
[music]
hey, hey.
Hey, hey, hey.
[music]
[music]
[music]
>> [music]
[music]
[music]
>> Hey. Hey. Hey.
[music]
[music]
[music]
>> [music]
[music]
[music]
[music]
[music]
[music]
[music]
[music]
>> Hey,
[music]
[music]
[music]
hey,
hey. Heat. Heat. [music]
[music]
[music]
[music]
>> [music]
[music]
[music]
>> Hey. Hey. Hey.
[music]
[music]
[music]
>> [music]
[music]
[music]
[music]
>> Heat. Heat.
[music]
[music]
Hey, hey, hey.
[music]
[music]
[music]
[music] Heat. Heat.
[music]
[music]
>> [music]
[music]
[music]
[music]
[music]
>> Hey, [music]
[music] hey, hey.
>> [music]
>> Heat. Heat.
[music]
>> [music]
[music]
[music]
[music]
[music]
[music]
[music]
[music]
[music]
[music]
>> Hey, hey, hey. [music]
[music]
[music]
[music]
>> [music]
[music]
[music]
[music]
[music]
>> Hey, [music]
[music]
[music]
hey, hey.
>> [music]
[music]
[music]
[music]
>> Heat. Heat.
[music]
>> [music]
[music]
>> Heat. Heat.
[music]
>> [music]
[music]
[music]
[music]
>> Heat. Heat.
[music]
[music]
Hey,
[music]
[music]
[music]
hey, hey.
>> [music]
[music]
[music]
[music]
[music]
[music]
[music]
>> Hey, baby. Hey. Hey. Hey. [music]
[music]
[music]
>> [music]
[music]
[music]
[music]
[music]
[music]
[music]
>> Morning.
[music]
[music]
Thank you for being here.
Okay, let me
Central, Brazil.
[applause]
>> [applause]
>> Graham McGrath,
Bank of England. [applause]
system.
[applause]
[applause]
Thank you all. Good morning everyone.
Thank you for being here today. It is a
great pleasure for me to welcome our
distinguished panelists here today. But
before I want to thank the organizers
for the this special space here to to
discuss this this uh very interesting
subject for me and uh also to the highly
qualified audience we'll have today. My
objective here is to is to uh have the
opportunity to encourage
uh further research on the topic we'll
be discussing today. It's a more policy
oriented debate here but we expect that
uh further research can be uh developed
in this area. So just starting um I have
just a few words about the the CCIB. You
may be aware the counter cyclical
capital buffer was one of the key
responses in back in 2010 uh to the
global financial crisis. It was part of
the bezel tree accord in Brazil. uh we
introduced the the standard in 2015 and
today uh its rate is is set on a
quarterly basis by the COMF the
financial stability committee here at
the central bank.
Uh the CCB
um design was supposed to uh it is
designed the the the the capital buffer
is designed to to uh accumulate be be
accumulated be activated or increased as
you wish as cyclical risks uh increase
build up over the cycle and to be
released uh when in times of stress. uh
back then in 2010 in its original design
absent those risks absent uh signos of
uh imbalances uh the rate uh should
remain at the zero%
which is the lower bound of the range of
possible values for the CCB
and uh and uh we refer to that base rate
as the neutral rate in Brazil it is zero
Europe and uh more recently the
international experience
uh has strengthened the case for uh a
more um [snorts] proactive a more uh to
start accumulating this this releaseable
capital. You will see us regulators uh
calling the CCYB as the releasable
uh portion of capital, the res
releaseable capital, the cycle adjusted
capital so on so forth. uh and there was
this case to to start accumulating this
parcel before in normal times. And the
idea is to uh uh is that if you have uh
releaseable capital available when you
release and uh this this uh policy
proved to be effective in helping banks
absorb shocks and also sustain the
credit flow. So um but as I I mentioned
before the the original design implies
uh base rate of zero and uh there is a
kind of it introduced a kind of inertia
that we will discuss here and uh for
several reasons and uh because of this
there is this discussion of moving
towards a more proactive approach uh and
to having positive neutral rates uh uh
in our framework. So uh I will I will
move to invite Mara. Mara she has an
extensive experience a deep
understanding of different designs
different types of calibration and
implementation challenges in the CCIB at
the European level. So MA over to you.
Thank you.
>> [applause]
>> Um
well good morning to everyone and many
thanks for to the organizers for uh
inviting me to to this uh uh panel
today. Um I don't see the slides on the
on the screen but just to start well
I mean my name is Mara Pirroano I I work
at the European Central Bank and it will
be a pleasure today to talk to you about
the challenges of um a positive neutral
CC bringing out the European uh
perspective uh the usual
disclaimer applies is my my own views
and do not necessarily
reflect those of the uh of the ECB or
the Euro system. Well, I can just uh
quickly start before uh the slide come
up. So just um uh well to start again
with the what is a positive neutral CCYB
actually uh this was actually one of the
main challenges we we encountered when
starting approaching this topic and um
well what it is not is actually a a
structural use of a buffer that was
designed to be counteryclical.
um
but rather what it is is the early
activation of this countercyclical
buffer uh in an environment when
cyclical systemic risks uh are neither
subdued or nor elevated. So what is
called like a standard risk environment.
Um well at TCCB we uh we have strongly
supported uh these early activation um
approaches and indeed to avoid a bit at
the confusion with the semistructural
uh use we have um shifted uh um towards
naming it as early activation rather
than positive neutral which gives indeed
a bit more of a structural uh um
delineation to this buffer. [snorts] Um
so we have um conducted extensive work
um on this topic. on the slide that I
was supposed to show you, there are
several um links to to publications both
more policy related and academic
publications which uh if you're
interested um well you may have a look
um but most importantly um together with
the European systemic risk board we have
launched a work stream uh back in um uh
2024 before uh that has uh included more
than 20 members uh from national
authorities in the European economic
area. Um and this works has produced um
a joint report uh which was published in
January 25 and on which uh the insights
I will present to you are based. Um, now
for the next slide, I would really need
a visual because it's a visual slide.
[clears throat]
So, I'm not sure uh how to um
sorry about that. So, well, the slide is
about the well try to improvise a bit
here. [laughter]
[gasps]
Um it's about uh the current
implementation of of the positive
neutral CCB in the fantastic thank you
um in the European economic area. No,
that was actually the this one. So um so
at the moment we have um 18 countries
that have uh adopted such early
activation CCB approach. Uh and actually
already from this slide you can spot
some of the challenges. I will uh I will
delve uh deeper into later. So first of
all you see that some uh countries have
a little red dot there and this red dot
are the target positive neutral uh rate
they have announced while some some
other countries don't have those little
dots. Um another um another um um
feature you see is theogenity in the
rates that are applied across countries.
So the rate nowadays from um uh 2%
target rate to um a 0.5 target rate. Uh
so well one of the main challenges in
Europe is indeed the heterogenity.
Um so
the 13 countries that have introduced an
explicit target rate um did so because
they considered an explicit target to um
ensure predictability of the rate
transparency visa v the banking sector
and clarity of communication.
um um other five countries instead opted
for an early activation approach without
an explicit
uh target rate to remain a bit more
flexible and they didn't really see a
need to to have an explicit target and
well the heterogeneous calibrations come
not only for from a policy preferences
uh a across countries but also uh from
the different calibration methods for
these target positive neutral rates that
have been used. And well, in this table
you see that quite a variety of of um of
methods have been implemented by
national authorities. Um and um however
despite most um authorities rely on some
modeling framework ultimately e expert
judgment still plays uh a key role.
There is also some heterogeneity
regarding the objective uh for adopting
such uh early activation um approaches
which well sometimes actually they uh
they coexist
and the main motivations um regard
timing and and amount. So on the timing
side uh uh authorities mentioned that
this approach is very useful to ensure
the early activation of the CFCB. a a
timely activation um and due to for
example data lags that may delay uh the
uh identification of risks that will
then result in an activation of the
buffer and also a more gradual buildup
of the buffer. So the earlier you start
then and the more gradually you can
build it up and then can has can have um
um uh more limited consequences on the
banking sector. And regarding the amount
well uh the motivations
are mostly well ensuring a greater
availability of buffers that can be
released uh in the event of a shock and
that was a a a big lesson from the coid9
pandemic and uh increasing the banking
sector's um resilience against um a
wider spectrum of potentially large uh
shocks.
Um so well we said that it's a buffer
that one should activate when risk
another subdued nor elevated. So a
standard risk environment but a
challenge is okay what is a standard
risk environment. Well, um it is a
situation where risks are neither
excessively high nor excessively low and
one can best identify it as uh the late
recovery period after a downturn or the
early expansion phase. And so the
picture here shows how uh the buffer
would be intended to be used um over the
cycle with um well an early activation
um happening in a situation where uh the
economy is recovering, banking sector
balance sheet are also recovering and
well overall banking sector conditions
are favorable. Then well the buildup can
be gradual towards the target rate and
then a further buildup to say the peak
rate would be um envisage when risk
really become elevated. Uh well and also
challenging challenges happen when to to
assess um how to then release
the CCB while either remaining above the
target rate. So for example when uh
cyclical risks recede uh or releasing
the buffer completely when really uh
losses start to materialize or you have
high expectations the losses start to
materialize to allow uh uh banks to
continue lending.
Um another challenges that that was
mentioned by uh quite some uh
authorities in our world stream is uh
the lack of clarity in the international
and EU legal framework. So um the Basel
committee uh on banking supervision
expressed explicit support for the use
of a positive cycle neutral CCYB in a
newsletter. uh they published in October
2022 and they even published in in 24 a
range of practices in implementing um
positive CCB reviewing the experience of
of countries that have implemented it.
Um
however this is not yet enshrined
formally in in the Basel framework which
for some countries constitutes um a bit
of a challenge and specifically in the
EU uh some jurisdictions um lamented a
lot of a lack of clarity in the EU legal
framework regarding the possibility
to um implement a positive CC B rate
when risks are not yet elevated. Um and
that's basically due to different um
transpositions of European directives on
in international law. Um another
challenge that was extensively discussed
uh regarded overlaps with other macro
credential instruments and uh this is
perhaps a more um EU specific um element
as um I mean in the EU we have the
systemic risk buffer uh whose objective
is to um prevent and mitigate
macrocredential or systemic risks which
are not covered by the CCB or other
buffers such as uh the buffer for other
systemically important institutions and
global uh systemically important
institutions. Um so um conceptually this
kind of uh vague definition of the
systemic risk buffer open the door for
using it also for objectives that the
CCB could address like well a positive
neutral CCB would address namely um
increasing resilience against exogenous
shocks. So uh and and this uh uh
this actually resulted in several
countries calling for more clarity in
the definitions um of this different
instruments that would really
disentangle their their objectives and
therefore avoid overlaps. Um finally
many countries um discussed the
importance of communication. Uh so
communication is clean is key not only
when you want to introduce
um a positive neutral uh uh CCB uh
framework. Um because this helps
obviously uh fostering acceptance by uh
the banking sector and the public. Um
and and it also helps steering their
expectations
uh regarding uh the setting of the raid
and its intended path over time. But
it's also very important when for
example you
would release a CCB
um because then uh it's key to
communicate very clearly what is the
timing and intended path of
replenishment because this would
encourage banks to actually use the
release capital to support lending
during a downturn and otherwise they
would expect maybe the rate to be
replenished too early and then they
wouldn't really use the buffers. Um and
and actually many um authorities uh did
engage in extremely clear um
communication
um especially through saying press
releases the financial uh stability
reports and and policy strategy papers.
So well um to conclude well the ACBM
believes that uh well um early um CCB
activation uh is a very valid um
approach that actually enhances the
counterical properties of the CCB. Um
well maybe greater consistency in its
application uh across jurisdictions
would be beneficial to enhance
transparency, predictability and
crosscountry comparability and um and
also improved clarity in the
international framework and uh and EU
legislation would really help uh to
clarify the objective um and uh the
design and the release mechanisms. Uh so
to really foster bit more harmonized use
which would also help uh the banking
sector to understand and and finally
clear communication is really an
essential element that has to be really
taken into account when thinking about
uh well introducing a positive um
neutral CCB strategy. uh and uh and and
also throughout at the use of the
buffering time and with this I conclude.
Thank you very much. [applause]
Thank you Mara. This was really
comprehensive. Thank you very much. Um I
will invite now an Estrada please.
Anhill being a former director of
financial stability regulation and
resolution in the banko despa
uh he has dealt with all kinds of
financial stability issues in in the
country and also the CCB and I suppose
that anhill was uh had a key role in
implementing this new approach that
we've been discussing here the positive
rates for the CCIB in the the previous
years. So, thank you. The floor is
yours, please.
>> So, thank you. Thank you very much.
Good morning, everybody, and thank you
for inviting me to to to to share with
you the change in the framework that we
use in order to activate and release the
contraical capital buffer that basically
moves from a situation in which we only
activate the buffer when risk systemic
risks were elevated. I don't know
>> when R were elevated uh to other
situation in which we activated also the
buffer when as Mar has said risk are at
at standard level. Thank you very much.
Thank you. So let me first uh move to
the
something I know that
>> I don't know what happened.
Thank you. So let me first uh try to
explain you why we decided to change the
frame. There were both uh factors that I
think affect to all of us and also
factors that were specific of the of
Spain. uh the common factors the covid
pandemic uh we arrived to the covid
pandemic without no macro potential
space with the contrayclical
capital buffer uh in zero basically
because we didn't observe the Spanish
economy the Spanish banking sector this
accumulation of systemic risk but the
thing is that the co covid arrived and
one tool that could be used in order to
mitigate the impact of the covid Um yes
a mitigation not a very hard mitigation
but a a mitigation of the COVID was not
there. Then we also have a lot of
empirical research in this period with
some countries that have activated the
contraical capital buffer and released
the contraical capital buffer and it's
very clear that activating the
contrayical capital buffer has some cost
and it has a lot of benefits when you
release the buffer. So the idea is to
minimize the cost and uh maximize the
the benefits. So one thing that you can
do in order to minimize the cost is
activating the buffer. Mala has said in
the early times in order to have time
enough in order to to to charge the
buffer and share the the the cause in
different and the other idea was that
super national organization were
advising to use this tool. For example,
in the the European case, the ESRB that
is a microproential policy, the micropro
macropential policy
um organization
they were advising to use this tool and
also the IMF was also supporting this
tool. Specific elements of the Spanish
economy were basically Spanish economy,
the Spanish banking sector suffered a
lot during the glo global financial
crisis. They accumulated a lot of
non-performing loans for close asset in
the balance sheet. So it was considered
by that time a priority to make a
cleaning of this of this balance sheet
and the cleaning as you know perfectly
well implies a lot of provisions.
Provision does that reduce the benefits.
So it was considered at that time that
by that time that uh increasing capital
requirements could be very harmful for
the economy as a whole. So once that
this clean sheet clean sheet was
finished that was basically after the
pandemic it was time to change our
approach and increase the resiling of
the of the bands basically using the
capital now not the promises and the
second element a very important element
is the volatility of our business cycle
that is very high compared to other
countries in in Europe and also the
correlation of the uh business cycle and
the financial cycle. So it's important
to have tools that help the authorities
to stabilize the economies and this is
very important for countries like Spain
where the other macro macro macro tools
stabilization tools for example monet
monetary policy is in the hands of the
European Central Bank. So it's not
designed specifically for Spain but for
for the Euro the Euro area as a whole
and pol fiscal policy has uh some uh
constraints taking into account the the
level of the of the debt. So first hard
that we find found when we tried to
change the framework uh was pointed by
MA uh the legal the legal hard because
in the first level in our first level
legislation of the European Union there
was a clear uh statement saying that the
buffer can can only be activated when
when there are risk and a nice
interpretation of this of this statement
is that there are there are only two
state two states of the nature in the
world risk and no risk. So if there are
no risk you cannot activate but when you
combine the financial cycle when with
the business cycle uh it's clear that
other possibilities could arise and one
is this h this situation of a standard
cyclical risk. So that was the first
interpretation that should be should be
changed in order to activate the buffer.
The second element the very important
role that the vessel gap has in the
activation of the buffer. Basel gap is
basically the uh credit credit to the
non- financial private sector divided on
GDP. The deviation from the trend trend
estimates estimated with a with a
filter. uh what is the problem with with
this with this uh indicator is the is
the what you have in the chart when the
pandemic came and the GDP declined
substantially this indicator increases
substantially indicating a activation of
the buffer precisely when the economy
was in a in a substantial decline. So
the use of the buffer following only
this indicator would have been recycle
recyclical no countercyclical.
Fortunately the legislation al also
allows you to use more indicators. So in
the change of when we change our frame
framework basically we have to take two
decisions. first how to um how to decide
where the the economy was in in which
phase and which state of the nature was
the was the economy and in order to do
that we develop a system or panel of
indicators. These indicators include
basically all the elements that we
consider that should be taken into
account. First macroeconomic indicators
because the tool should be
counteryclical
shouldn't be procyclical. So you you
need to have this kind of indicators
micro financial because is the financial
cycle where what we are trying to tame
indicators from the financial markets
because they have very good properties
as early warning indicators and also
indicators from the banking sector
because we have also a lot of evidence
that uh says that um to minimize the
cause of the of the activation of the
buffer the banks should be in when the
banks are in a better situation the cost
of the activating the the buffer are
lower. So it's important to take into
account the situation of the banks when
you activate the buffer. Second element
and Mara was was presenting this also
very clear is very important the
predictability all these uh all these
indicators should be reproducible by all
the people that is going to suffer the
the buffers the banks and all the all
the all the people and they they should
have a very clear evidence or a very
clear guidance on how the buffer is
going to be used. As you can see the
chart is very similar to the chart that
has been shown by Mara and it's very
important to um to to to communicate to
the to the to the people that is working
in the market how is going to be to be
to be used. In our case we have a let's
say a forward guidance that is dependent
on time but is also depending depending
on the evolution of the data. And the
final element in order to in order to to
determine to to operationalize all this
new framework is determining the rate
that is adequate for the situation of
standard cyclical risk. What we did in
this case was basically to use the tools
that we have in order to to to do the
stress test but in a different way
because usually the stress test are
used. You have the actual situation the
current situation of the of the economy.
you define a a scenario let's say a a a
most likely scenario and then a scenario
that is totally stressed that is
dependent on the most likely scenario.
In this case what we did is to reproduce
the economy in a situation in which
there were not imbalances
there were not accumulation of systemic
risk. the situation in which we could
say that is a standard cyclical level of
risk and then we shock the economy with
different shocks shocks domestic, real,
financial, external and we shock the
economy. This this these shocks have a
different intensity in order to have
different implication and we calculated
the the capital consumption in these
different situations. As you can see, we
have three kinds of three kinds of
intensities, mild, medium, and severe.
And in the end, uh it was decided by our
governing council that the situation
that should be covered was the one that
is all the shocks affecting the economy
and uh the intensity mild. Why? Why
mild? probably the explanation is
because the probability of having a
selfcorrection of the economy in a
situation of a standard level of risk is
is reduced. In the end, we decided to
put the uh the the buffer in 1% and you
can see the difference between the 04
and the 1%. It is due to the activity
that the Spanish banking system has
overseas over overseas. In Spain
approximately half half% of the activity
is uh overseas. So for a capital
consumption of 04 we need to put a a
contrayical capital buffer at 1%.
because the countercyclical capital
buffer this is a little bit technical
but I think it's important to to
communicate is an average of the counter
cyclical capital buffers of that all the
bands have depending on the where they
have their their position in the
different countries so and that's all
for from my side thank you
[applause]
>> thank you a very crucial show insights
and especially this last slide. Uh so
connecting the dots I will invite Gan uh
to the floor. Thank you Grean
from the Bank of England. Bank of
England I don't know if you know but it
is one of the forerunners of this idea
of positive rates. Will has an extensive
uh experience dealing with stress
testing and stress testing is one of the
main methodologies used to to not only
to understand resilience but also to
calibrate the buffer and uh I think that
you also have uh vivid memories of some
disruptive events he will mention here
for example Brexit and covid and this is
very interesting experience that we we
can assess. Thank you, Gorilla.
>> Thank you.
[applause]
How do I
>> um so thank you very much. So I'll just
say a few uh words about how we've used
our experience of using the the positive
neutral CCYB and I think happily it will
coincide with a lot of the ideas that
Mara and Anel have already outlined. Um,
so thank you very much. Um, so
this uh, so just to say we've been
active users of the CCB since it was
first introduced and you can see this
chart here shows the level of the CCB in
the UK which looks a little bit like the
kinds of charts that that Mara was
showing uh, earlier. And I'll come back
and talk about some of the specific
moves that we've made
um in a few minutes. But just to say
really by using this the CCB in the UK,
we're trying to achieve two things. So
the first is making sure that the
banking system has enough capital to
absorb the kinds of losses that it might
experience in a severe financial or
economic outcome. And the second thing
is making sure that the banking system
will actually use those buffers when the
shocks crystallize. Um, so what we what
we're trying to do is avoid the kind of
really costly deleveraging that was seen
during the financial crisis. And what we
like to say is that if a a shock hits,
we want banks to be making decisions
about who they lend to in terms of
households and businesses based on the
balance sheets of those households and
businesses and not based on worries they
might have about their own balance
sheet.
Um so it's set by the financial policy
committee which is a committee within
the bank of England. It's chaired by the
governor and uh it has many overlapping
members with our monetary policy
committee and also our supervisory body
as well all of which are under one roof
which really helps uh coordination.
So in uh 2019 we increased the neutral
rate of the UK CCYB to 2%.
And uh there was really a few things
behind this. I think this was probably
one of the earlier moves towards 2% in
terms of a neutral rate. Um and the
there was two two parts to the thinking
behind this. The first was that evidence
showed that in order to be able to
absorb um the kind of shock that could
hit at the top the impact of a shock
that could hit at the top of a of the
financial cycle, the CCYB would have
needed to have been between three and a
half and 5%. Uh and that would have
meant that there was enough buffers in
this banking system to absorb those
losses without banks going into their
minima.
And the second was that um most looking
back at 2007
uh most indicators didn't really
indicate that there was a need to build
resilience until quite late on. So this
chart is quite complicated, but the
yellow sway shows what most indicators
of financial stability were pointing
towards. And you can see that most of
them didn't really become elevated until
late 2004, early 2005. and that at by
that point it would have been where it
would have been too late for the FPC to
gradually raise the CCYB and if they'd
wanted to get to three and a half or 5%
they would have needed to make some very
sharp moves in the CCB which wouldn't
have been good because it would have had
a uh an impact on the supply of credit.
So by contrast with a higher starting
neutral rate of the CCYB at 2% it gives
the committee more chance of being able
to increase the CCB rate gradually as
risks build and be able to get to where
it needs to be to absorb the kind of
impact of shocks that could hit at the
top of the financial cycle.
Um so we increased the rate to 2% in
2019 and that had two uh two real
benefits. The first is it gave banks
more releasable capital buffers to be
able to absorb losses without reducing
lending. And the second is that as I
said it facilitates a gradual approach
to raising the CCB as the cycle becomes
elevated. And um we've really tried to
communicate that we will take a gradual
and predictable approach to increasing
the CCB because that has less impact in
terms of the day-to-day uh lending.
The other important feature of this was
when we did increase the neutral rate
from 1% to 2%. We didn't want to change
the total amount of capital in the
banking system or the total amount of
loss absorbing requirements. So at the
same time that the neutral rate of the
CCIB was increased from 1% to 2% there
was an offsetting change made to reduce
pillar 2A requirements which helped us
get uh over the line with that and um
reassure the banking system that this
wasn't a way of increasing the
structural amount of capital in the
system.
So just to say a bit more about uh how
we've used the CCYB rate actively um and
in particular when we've released it. So
we've released the CCYB on two
occasions. Uh the first is just uh here
in in 2016 you can see we just managed
to increase the CCB to half a percent
and had announced it was going to go to
1% but then um the UK voted to leave the
EU. Um which was a little bit of a
surprise. Um I'm Irish by the way so for
me it was quite a big shock. Um although
I did go on to lead the Brexit division
in the bank which my wife never never
forgave me for. Um but uh we we reduced
the CCYB then to back to zero and the
the narrative we gave was that in times
of uncertainty
we wanted to make sure that banks had
the confidence to continue lending to
creditw worthy businesses and house and
households.
Um but then later about a year or so
later it was clear that the risk
environment hadn't deteriorated as much
as we had expected or as much as we had
feared. And so in 2017 um the committee
was of the view that the the risk
environment had returned to standard and
so it increased the CCYB back to its
neutral rate of 1% over the course of
2017 and 2018. Then there was that
increase to 2% um in 2019 which I just
talked you through and again we had just
managed to get there uh when the co
pandemic uh hit and immediately the FPC
was able to reduce the CCB to 0%. And
again that really helped in terms of
sending a strong narrative that the
banking system would have the capacity
to continue to support lending
households and lending to households and
businesses through the COVID pandemic.
Um and that that that pandemic release
gave us a really nice um uh example to
be able to study about the impact of the
CCB. And there's been lots of research
done on this. I've just quoted the study
that was done by some Bank of England
staff here, but I know that there was
work done by the BIS and also work on
the EU and [snorts] Hong Kong
experiences as well. But the key finding
was that um releasing the buffers at the
start of the pandemic showed that banks
who who were more who benefited more
from that cut in the CCIB maintained
higher lending and lower loan prices and
they also continued more lending to
riskier borrowers. Um, and this is a
really really key finding for us because
that's exactly what we're trying to
achieve through macroecial policy. Going
back to what I said at the start, we
want banks to continue lending to
creditworthy households and businesses
through a downturn and avoid the kind of
reduction in credit supply that was seen
in 20078.
And then just the last thing I'd say is
that the banking system has uh really
been positive about the benefits of
releasability. So they feel that
releasable buffers are just more usable
than non-releasable buffers. And one of
the things they talk to us a lot about
is uh maximum distributable distri
distri uh MDAs which are uh trigger
points at which when their capital falls
they're restricted on the amount that
they can distribute through profits and
uh dividends and that kind of thing. And
by reducing the CCB rate, it reduces the
level of capital at which those things
are triggered and they're very important
uh for the banks themselves because
they're they're important for their
shareholders and that kind of thing. So
that's one of the things that they've
been very positive about in terms of the
CCB and um and the FPC's use of it. And
I think the fact that the FBC has
released it on two occasions has really
helped reinforce credibility that this
isn't just a one-way street that's going
to be used to build capital. it will be
released when it needs to be as well.
Um, and then yeah, just a final word
which goes back to some of the the
points uh that that Mara and Anel made
about h the importance of communication.
So, because we've been using it for um
you know 10 10 or more years now we've
got a wellestablished framework for
using the using the CCB. We've got a
policy statement which we updated a few
years ago on our website that outlines
the kinds of things that the FPC take
into account when they're setting the
CCIB. Uh some of the conditions under
which they might cut it and some of the
conditions under which they might raise
it to above neutral as well. Um and
there's a two-step process that we
follow. So very similar actually to to
what Angahel was uh outlining for Spain.
We assess the level of financial
vulnerabilities and uh take a view on
where we are in the financial cycle. And
then we assess the ability of the
banking system with to withstand shocks
without uh without needing to resort to
an unwarranted reduction in lending to
households and businesses. And I think
this is one slight difference. Um we
look at a we don't have a single
indicator or a single mechanical uh
thing that we look at. So we look at a a
wide range. We look at macro indicators
across the financial cycle. Um we look
at trends in lending and how that
compares to how the economic outlook is
evolving and we use intelligence from
banks and businesses as well as to what
their experience is. We do also use the
stress test. That's a really important
that is a really important input and um
we also tend to vary the severity of our
stress test. So at the top of the
financial cycle, we'll have a more
severe scenario and then we'll look and
see what the results are imply about the
amount of capital that the banking
system needs uh to have.
And then lastly, just to say um we do we
do try and respond to some of the
feedback feedback that we get from the
banking sector. Um and two two things
that we're taking forward as part of our
capital review that are relevant is one
is how to enhance the usability of
regulatory buffers more generally. So I
mentioned earlier that the banks talk a
lot about uh impediments to using other
capital buffers um such as the COB or
the systemic buffers and two things that
they raise there are that those those
buffers aren't releasable. So they would
prefer if they were more like the CCYB
and the MDA point that I mentioned
earlier. And then the last thing uh that
we get a lot of feedback from from our
banking sector is about the interaction
of the CCB with other requirements that
are also calibrated on domestic
exposures. So they um raised the point
that both the CCYB and our domestic
systemic buffer and also uh credit
concentration risk requirement are all
calibrated on the level of domestic
exposures. Um so they've asked us to
look and see if there's any overlap
there which we're which we're doing. Um,
but all that to say that I think it's
it's uh the the I guess I'd sum up by
saying that the UK's experience of
having a positive neutral CCYB rate has
been really positive and we've really
felt that it's really helped uh enhance
our macro credential framework and and
help ensure that the banking system can
support the real economy through uh bad
times as well as good. And a lot of the
things that we've already talked about
such as being really transparent about
how we use it and responding to feedback
as well have been really important for
trying to enhance the credibility. Um so
that's it. And then I think the one
thing I'd just also mention is we did a
lot of research which I haven't listed
here but over the years there's been
some really productive research done. So
um I'm happy to send on a few links to
papers and things that have been done by
Bank of England staff but I think
there's probably a lot more that can be
done on that as well. So that's it.
Thank you very much.
>> Thank you very much. [applause]
It's very inspiring.
Um, so now Yonour, thank you for being
here. I will invite you to to the for
Thank you, Miku. Yonour, he he's from
the Swedish Financial Supervisory
Authority, which shares responsibility
with the risk bank for micropedential
monitoring and also the CCIB.
uh setting
uh Yonour uh faced the challenge of uh
of uh Do you want my help?
>> Stop a bit.
>> Perfect.
>> Thank you. Thank you very much.
>> I thank you. I would just say that you
you have extensive experience dealing
with in a highly open economy like
Sweden and also an active user of the
CCIB as a microcredential tool. So thank
you.
>> Go ahead.
>> Thank you very much.
[applause]
>> So So actually uh I've only lived in
Sweden for two years now. uh been in in
this uh uh in this role for for two
years but but prior to that uh I
actually during the initi initiating
phase of of counter technical purpose I
was uh uh directing the uh deputy
director of the financial supervisory
authority in Iceland an even more uh
remote country but also in the in the
European family. So so uh but very
similar experiences of when we
introduced this in the first place. and
we've had some very in very excellent
contributions so far. So, so I will try
to be as uh a little repet repetit
repetitive as possible but but but um of
course it's a little bit dangerous. I I
thought I would start
the value of of my my is speaking here
would be to give you like this the story
of Sweden and how that really fits into
all of this. And this picture is is is
basically uh uh fraudulent on the basis
we are calculating the buffer guide
retroactively. This of course was not
even an idea in the early 90s that this
picture goes back to. But I wanted to
illustrate that Sweden is a country that
has experienced quite significant
financial crisis episodes in in in over
the years. Of course, the the great
financial crisis everybody knows. Uh in
fact, Sweden was mildly affected by that
by most standards. Uh and that that
really uh affected how how Sweden then
started developing the the counter
before that. But in the early 90s, uh
Sweden had a very serious financial
crisis with several banks being being uh
rescued by by authorities. So definitely
uh there's no hope that Sweden has has
found a cure for financial crisis. Uh
but but there's also another point of
bringing out this this picture of of
basically the uh the credit gap and and
the buffer guide pro provided by the
bash of the court in that uh while it it
is is is is quite indicative at at times
in other times it really does not really
suggest that that this buffer should be
used and if you look at the end of the
the figure that is basically the period
that we we we are talking about.
So uh but but when we can we can think
about so this is a picture of how the
buffer has been been uh set in in in
Sweden from from so and the dates
referred to the to the date of of the
implementation not the announcement as
in all all of these countries the
announcement is usually one year in in
advance when it comes to activating and
increasing. So uh very much from the
outset of of the new uh rules on the
European level uh Sweden uh started
applica applying this at the 1% level uh
very much with with respect to the the
uh say traditional uh metrics that were
designed at the outset but but continued
gradually in increasing the the buffer
rate for uh well very much up until
2020. 20 but with different arguments uh
as we went uh along. So while it was the
credit growth that was the main argument
to start with, other arguments are more
prolific or important for the decisions
in the later periods. And this has to do
with underlying vulnerabilities of the
country in question. And and in the
Swedish case, this is primarily uh a
high level of indebtedness both in in in
for households and and corporates. and
uh also that uh really um housing prices
had not slumped as as they did in many
countries during the great financial
crisis. So continued high uh prices,
continued high uh vulnerabilities and on
top of that extremely large exposures in
the banks towards these particular uh
vulnerabilities in in related to real
estate. So more than twothirds of all
exposures of banks are are real estate
in some some sense or another. So the
banking sector uh is is particularly
vulnerable. Uh and this is of course a
key reason why why this was early uh
applied in in Sweden.
So uh another uh chapter in the story is
has already been mentioned during the
the co pandemic uh Europe as a whole
applied this this this this uh this uh
option of releasing the buffer and and
uh few countries had uh made as big a
splash in this as Sweden uh as it had
had ma
been able to to raised the buffer up to
two and a half% uh uh prior to that
event
and we can very generally uh describe
that that experience as as as positive.
This really gave banks uh the much
better leeway to to uh accommodate a
very very tricky situation uh at this
time. Of course, the main uh objective
was to uh uh uh let's say counterbalance
the the extreme uncertainty that was in
in in in place at the time. It was
definitely a count preemptive move
because even though the risk was that uh
that that that credit quality would
deteriorate dramatically and we would
have more losses in the banks uh this
was of course done based on the
expectations that that might might take
place.
And let's move on to the the third phase
which is then uh the topic at hand which
is the the the positive neutral. uh now
and and as I have explained the the
early adoption of the buffer is really u
um has changed character over time from
being a very formal output gap oriented
to let's say a more holistic uh view of
of what are v vulnerabilities have and
and we we can in some ways say that uh
this this idea of positive neutral had
already started in in an informal way in
in Sweden before the the pandemic but
[snorts] it was formalized uh right it
it after after the pandemic very well
communicated I really uh support uh that
the emphasis from from my fellow
speakers uh and then uh the the the
buffer was raised in in two two steps uh
to to from one% to two 2%
uh very much at a speed that was uh
fitting for the situation in in the
banks Of course losses were not as
severe as as were feared and the the
headroom in the banking system was quite
significant to to uh to actually
accommodate this this increase. This is
did not have any serious impact on on on
the u uh ability of banks uh to to to
keep lending.
So, so the main arguments uh that were
that were put forward uh in Sweden very
carefully uh communicated uh for for
having a positive neutral capital buffer
uh were that that that
the ability to be able to release
capital buffers in a future event is is
is is most important. uh and and of
course we have have the lesson of co to
to really support that that fact but but
also because uh economic stress
financial stress is very hard to predict
and and if you uh are able to come up
with a model that is very good in in in
defining when a financial crisis is
likely to hit chances are that that you
are basing your estimates on data that
is that is quite old and and you will
almost in all instances is only be able
to very very carefully and and and and
let's say uh granularly calibrate the
the buffer rate uh
except way way too late but perhaps when
when the whole event has has has been
been over
uh and so so now Sweden has this uh this
level of 2% as as kind of what we would
call a a normal rate uh when there's
neither elevated uh risk in in the
financial system nor is there uh any any
any let's say crisis ongoing. I'll skip
for time sake very quickly over this. I
I just wanted to make uh uh one point
about the methodology about how how this
is calibrated and my main message is
that this is not a a a symmetric
process. there are quite different
considerations to to make when you are
on the buildup of of the cycle than than
when you are considering to to release
it. So, so in our methodology we we have
like let's say two sets of indicators
that that are are are and you can even
argue that there should be more because
releasing uh from from different levels
whether you have reached the the
positive neutral or not is also a
consideration.
Uh
I I I I I want to make one one
additional point uh uh that that is uh
may not seem that related to positive
neutral but but it is really in the end
and that is the usability of it has been
uh raised before but but this is a quite
a become a quite complicated issue uh
worldwide but I would say particularly
in Europe where the implementation of of
pastor 3 is a
evolved uh and uh so we have done an
analysis on our particular banks in
Sweden and uh what this picture on on
the left tells you is is well the yellow
lines would indicate what individual
banks would have as headroom simply
looking at their their C1 or or or
common equity numbers. But if you take
into consideration that there are other
restrictions also applying they might be
from the liquidity uh rate so sorry the
leverage ratio they might also uh be be
from from uh rating agencies other types
of of of of concerns uh and also most
importantly in our case uh the the
resolution recovery system. So there are
certain requirements set forward uh so
so so so banks uh let's say
significantly large banks can be
resolved in a in a cost-effective and in
a smooth way and these requirements are
often a more binding than the actual uh
capital buffers. So, so the
maximum distributable amount I made it
[laughter]
>> very
may not be sort of the the relevant
issue here. Then there's also a a a
concern with with that way for that has
already been raised. It can contain a
stigma fact to it. uh the experience
that that we have at least is that uh
with with the counteryclical buffer. Its
release uh was was highly uh appreciated
as as a as a and used in the intended
way albeit with a with of course an an
added let's say recommendation to limit
all kind of distributions from from
banks during a particular period because
of course the way it turns out that
could have could have been been been the
result. So, so on on the right hand side
of the picture we see at least one one
of these banks that is really uh you can
see what kind of restrictions are the
binding restrictions uh in case of of
any any any losses occurring and they
are as you can say see a little bit all
over the place which is a concern but in
that uh having a a positive neutral
counteryclical capital buffer is really
a plus because that is is is the most
releasable buffer. You can you can you
can say it's something that has already
accepted as something that banks can use
in in in times of trouble and and just
do not really really affect uh their
their their trustworthiness their their
uh the views of of rating agencies the
the view of of of their their bond
holders etc. Uh so that is basically my
my me main message that I wanted to
convey here uh that u we we we don't
look at this as as symmetrical it's it's
asymmetric and it's very important that
that you have
usable buffers that can really be
applied in times of of crisis. Thank you
very much. [applause]
Thank you all for your presentations.
Thank you the panelists and uh I think
this helped uh bring everyone to the
same page. Uh maybe not everyone here
has a deep uh uh contact with the the
capital framework and the the
counteryclical buffer. So uh this also
raised uh very interesting reflections
and questions. So I'll jump I will go
straight to one first question in the
benefit of time and then I'll move to
the audience to to additional questions.
And a key feature that I I I see in all
of your presentations refer to the
uh starting uh accumulating the capital
the the release the releaseable portion
of capital uh in normal times in
standard risk environment
um
and also in a gradual manner. And this
uh for me I saw some graphs there and uh
this a common rational for that as I I
was researching about is that um it is
less costly for banks to raise capital
in good times uh in in benite micro
financial conditions and uh compared to
um having to increase fastly or to not
having buffers accumulated when the
shots arrives. And a question that comes
to my mind is okay but what uh uh
information indicators I I know that you
mentioned some but what additional
indicators
uh to inform our decision- making
process our our analysis uh that could
uh considering this this argument of
having uh a costbenefit
uh assessment in good times. what kind
of indicators what kind of information
or how should we define this standard
risk environment uh it is for you all I
don't know who wants to start but uh who
wants to react first
is it tough okay go ahead go ahead
>> yes thank you Fernand can you hear me
[clears throat] okay this is a very
relevant question of course it's about
uh okay when is the right timing to
introduce uh well to start building up
uh uh early in the cycle and then well
obviously timing has a uh very high
relevance because if you choose to build
it up too early then uh maybe the
economy and the banking sector haven't
really fully recovered from a downturn
and then you really risk being proyical
and and and constraining that that
recovery.
Uh so I mean in our experience what what
is useful to look at are both uh
macroeconomic indicators like to have an
idea of uh
how your recovery is well your economy
is recovering from from a downturn. uh
but also uh the conditions in the
banking sector and and most uh well the
two most relevant that that have emerged
also in in academic studies that
colleagues at the ACB have have uh have
done are uh profitability
and capital headroom because a
profitable banking sector um uh means
that if you increase requirements then
they can easily meet them with return or
earnings uh and and that's obviously
less costly and and capital headroom
like meaning the the the headroom well
the voluntary buffers they hold uh on
top of requirements then if there are
ample it means that they can easily
convert those those uh management
buffers into requirements without
needing to issue new new expensive
equity.
>> We are changing the nature from uh
manageable management capital to
release.
>> Exactly. Basically what they have they
just lock in into a buffer without
needing to raising new expensive equity.
And and that also means that the uh that
the
the impact of lending will be much
reduced over or even nil. And I mean the
exp the experience we have uh we have
had in the in the Euro area is that uh
well several countries have introduced
uh
a positive neutral rate framework uh or
uh say simply increase
buffers
even when uh monetary policy
was tightening. So in the titing period
that has started in 22 uh basically in
the space of one year uh the ECB's uh um
marginal refinancing rate rose from 0 to
4% which is quite a lot. uh but in that
very same period many countries I think
six if I remember correctly did increase
their their CCB
and the reason they could do that
without being procyclical
was there the banking sector was
exceptionally profitable due to the
higher interest rate and they already
held quite some voluntary buffers. So
then
>> these are essential conditions.
>> Thank you.
>> Thanks.
>> No, I basically agree because having
said I think it's very important to
distinguish in the different states of
the natural natural way for for example
when you are in a standard level of risk
macro macro variables and banking
variables I think that are the most
important. Erh we have empirical
evidence showing that when the situation
of the bank is good
the cause of increasing buffers are are
very reduced also when the output gap
defined as you want is positive the
situation is also is the cost are also
uh reduced but when you are in a
situation in which you are accumulating
systemic risk I think that the financial
variables credit the traditional
dominate. Yeah, absolutely. And in those
cases probably if you are accumulating
your credit is increasing by 20%. For
example, I think that you have to add in
any case independently of this
macroeconomic situation and even on the
situation of the banks and then that
there could happen situation in which
you need to add also for example uh
increases in the housing prices that are
not the result of supply of a relaxation
of the credit credit standards. those
situations as the reduction in housing
prices the main impact that is going to
have on banks is through the
macroeconomy. I think that the capital
buffers could be also a very good
response.
>> Thank you. So I will open the floor to
to questions of please uh yeah thank you
say your name and your institution and
the question.
Hi everyone. Um, Douglas here from the
Bank Central Brazil. And before I even
ask my question, just have to say, wow,
what a great panel. Thank you so all so
much for all these great insights. My
hand almost hurts of how many notes I
took. So uh really useful uh insights
and uh so grilling I think you put very
well uh in you highlight a number of
times in your presentation that a major
goal of the positive neutral CCYB
is to allow banks to preserve the
capability of supply credit to the real
economy other financial services to the
real economy when a major shock hits and
regardless of course whether the shock
is financial or not financial as co it
was
uh but for countries like Brazil and
other countries which have a 0% CCIB
you know uh one concern that I hear from
some market participants from some
commentators is that the liftoff from
zero to an eventual positive target or
to a positive rate the implementation
itself the liftoff could unduly
constrain credit supply in normal times.
So I would love to hear from you know
people in panelists
what were your experiences
from this implementation period if you
look back how was credit supply
responding and how did that uh uh how
did that relate to industries accenti
expectations you know when when it was
first being discussed uh did it
correspond to their worst fears or did
it or or was it not as constrained in
reality uh when it actually took off
from zero to a positive rate. Uh this
will be very helpful to for us to to
consider as well. Thank you all so much.
>> One more.
>> Hi, I am Francisco also from the central
bank of Brazil. I have a question um for
anyone in [clears throat] the panel uh
who might want to to take to share with
us. Uh and that is regarding uh the
eventual need or convenience of
coordination or adaptation of the
calibration of rates across
jurisdictions.
uh considering the possibility of uh
arbitrage
for capital requirements no across
considering the international banks
considering that some multinational
banks might uh move they they they kept
their structure within the group you
know and and is that should that
possibility be taken into account when
uh the financial supervisor is
calibrating its uh neutral rate. Thank
you.
[clears throat]
>> One more. Okay.
>> Yeah. So very interesting panel. uh my
question is more about the release phase
that we did that was not so much
emphasized uh uh during the panel. So I
was wondering in your institutions
how moving from a neut from a neutral uh
to a positive neutral a counteryclical
buffer from from the zero to some
positive number has affected or at least
interacted with the discussions about
rules on restrictions of payments of
dividends during the release phase. how
these two issues have interacted
[snorts]
uh in your institutions. There seems to
be an academic debate about the
>> about the pros and cons of of the
restriction of payments of dividends in
a provoking sense recent discussion. So
I wonder how this has uh fed up or fed
up uh into your institutions. Thank you.
>> I can start. Um very good questions. Uh
I'm not I will not be able to answer
them completely. So that's good. We have
good panel here.
>> Question. Thank you.
>> Um so first about supply of credit in in
during a shock. I I I agree it's it's
it's it's the fundamental uh reason for
why this is in place. Uh the the risk of
of deepening uh financial crisis because
of banks being very restrictive and in
their lending uh and and and
exacerbating the the effects on on the
real economy. It's really a a
let's say a key key uh thing. Now when
it comes to the way to get there, what
you what you call the liftoff, uh this
has been a a key consideration. It was
also mentioned here earlier that that
when you when you whether it is in the
old style or moving to a a positive
neutral rate.
I I agree with with with Mara that that
the key consideration is that that that
banks are profitable uh and and the
headroom is is there. There are other
indicators that you might want to look
at as well for for instance uh how how
NPLs are etc. how the price of risk is
is uh and asset price levels. But but
definitely uh and this is also really
very much related to the the third
question uh when you want to build up a
positive neutral rate at least uh you
you want to do this in a in in a in a
cost-effective way for for the whole
whole system uh and and and it so
happened to be possible in in the case I
des described in some other cases due to
the uh the the uh let's say the headroom
that was left from from the crisis
because the crisis would not was not as
serious for for uh NPL's as as initially
uh considered. Now uh
regarding the uh the uh the u
jurisdiction issue uh and and this is a
case that that is very familiar to to uh
Sweden which is a Nordic country with a
lot of interaction with with with
neighboring banks from Norway and and
Finland and and Denmark. uh so so like
two out of uh five major banks operating
are foreign banks. So this is a key
issue how these capital buffers are are
set. Uh this is a what is in the
European lingo called reciprocity. So
there's a system where you where you uh
uh want the the capital buffers to apply
to to exposures irrespective of the
their sort of where the where the sort
of lender is coming from. But but where
the risk is is is located and and this
is sometimes quite hard and this is
particularly difficult when it comes to
the uh infamous systemic risk buffer
which is a complete European uh
invention uh and and we've had uh
episodes particularly from between
Norway and Finland and I don't want to
go into details here but but where where
authorities have not agreed on the
interpretation of what risks are covered
by a particular buffer
thereby led to uh inconsistencies in the
application of that. But but the uh
positive here is that the the counter
cyclical capital buffer uh in most cases
alleviates that. So the rep repress this
is another very hard word
>> reciproc reciprocation
um is is quite automatic. So uh in in at
least if you don't reach a level of over
two and a half% and and that is no one
has done that [clears throat] before uh
there are there are of course uh uh and
and and we I don't know if we will ever
see see that high levels. So, [snorts]
so uh so this uh I can I can surely say
is is is a reason why some countries
have chosen
that buffer rather than the systemic
risk risk buffer in in in some some some
occasions. Uh and and uh I assume a
similar system would would would apply
to to this region.
>> Can I just add on the the question about
raising the CCB for the first time? Um,
so I think in 2015 the increase in the
CCYB to 0.5% was really part of the post
financial crisis capital rebuild. So
it's in it's difficult to to pick out
what the specific impact of that was but
our experience generally was very
similar to what other panelists have
said whereby
um the academic research that we've done
finds that if you increase the uh CCYB
during times when banks are profitable
and in normal times it's very it's much
less costly and it's has very little
impact on lending. Um, and that's also
consistent with other findings where
we've we've found that raising the CCYB
in general um has little impact on the
level of of lending when it's done in
normal times and can be met through
retained earnings. Um, in 2021 it was a
bit different because the banks had had
the headroom so that you know the the
effect of the co pandemic hadn't been as
as bad as people feared. So banks still
had a lot of headroom that they were
able to use and at that point we were
able to raise the CCYB um relatively
quickly back to its 2% and it had very
little impact on on the supply of
credit. Um and I think
when yeah I I think currently as well um
we sometimes hear from banks uh they
complain that the level of the CCB in
the UK is is higher than other
jurisdictions. Um we have two things to
say to them there. The first is that
they we need to remind them that we
reduce the pillar 2A component of the
capital stack at the same time. But
secondly then when we look at the the
amount of lending that's going on in the
UK and the profitability of of the UK
banking system, it sort of doesn't ring
true with with them being constrained by
capital. So we kind of make sure we do
an independent check. And then just on
the the payouts question, um just to
pick up on that one, I think there's
probably three elements. So first is if
we um if we reduce the CCYB the
intention is that it's above it that
that capital can be used to support
lending if if banks incur losses. So we
wouldn't want intuitively you wouldn't
want banks to just pay it all pay all
that out and reduce the capital um
through distributions to shareholders
etc. So it sort of makes sense that you
wouldn't want them to increase dividends
beyond what they would ordinarily be
doing. um but there's still scope within
that for them to pay out dividends and
to pay bonuses etc. so long as it's not
eating into their capital buffer. I
think then there's a separate tool about
whether you want to completely restrict
dividends and set them to zero which I
think of personally is just a different
step that you might want to take which
is that bit more extreme than uh just uh
simply saying don't increase your your
dividends and then just to come back to
the MDAs I think even um even putting
those two things aside there is a
benefit for banks in that it's kind of a
technical thing but if they were to go
in go into their combined buffer, they
wouldn't be able to pay out the coupons
on their AT1, which is a instrument they
have issued and just releasing the CCIB
allows them to continue doing that,
which is something that that they find
very helpful.
Yes, if I may complement [clears throat]
what uh well just said u regarding
dividend uh dividend restrictions. Well,
uh when when uh the pandemic hit, uh
then uh the ECB actually recommended
banks not to distribute dividends. Like
there was a big credential relief uh
accompanied with a recommendation not to
not to um distribute dividends. And uh
there was actually um a paper uh
assessing the impact of that uh policy
on lending and um [snorts] the paper
find um a positive impact on lending
specifically uh to the economic sectors
that were mostly hit by the pandemic. So
I think this this actually shows that uh
well it's an effective policy to combine
uh with a relief to ensure that really
credit uh goes to uh the real economy
and and money does not go to the pocket
of
investors which is not really the
primary objective of a capital uh
release. Uh and then um well there was a
question uh regarding um uh regarding uh
um coordination across jurisdictions
I mean uh the CCAB has a mandatory
reciprocity until 2.5%
so until that rate uh there's automatic
reciprocity so a level playing field is
guaranteed h If the question is
should a positive neutral rate be
reciprocated? Well, uh we have analyzed
that question in the context of this
European work stream and uh the answer
was uh yes simply because
we're not dealing with a new instrument
is the same instrument. So the same
rules that apply to a say a regular use
of the CCB
would apply to to its early activation.
So
just to clarify that uh well we're
really talking about one instrument and
and and the same design rules apply
>> no matter
whether you choose to activate it early
or or or you use a more traditional
approach
>> just to finish. I know everybody wants
to to have lunch. [gasps]
So, thank you very much for your
contributions here. the today's debate
uh shows that uh the CCB role is
evolving how it is evolving and um uh
for me it's a paradigm from um uh
passively react to uh visible
uh kakasha
[laughter]
vulnerabilities to a new approach which
is is the same instrument but a new
approach uh where uh we are uh concerned
about uh building progressively building
uh resilience over time in a
cost-effective way uh in order to to
deal with uh uh increasing shockprone
world nowadays. And one additional thing
it's a I think uh it's a powerful
insight to have um to understand and
this is food for research I think um
that the cost of building capital is uh
scenario dependent for me it's very
interesting and uh how to investigate
this issue and and to understand this
and uh there is this insight for me that
the dynamic capital can be a cost
effective way of deal with financial
stability issues and it is gaining and
more important nowaday in the future. So
thank you very much for your attention
and thank you all. [applause]
We can have a lunch
Thank you to our speakers. We may now
return to their cities.
No.
opinion.
Portuguese
>> [applause]
>> Liquidity or wealth, consumption, debt,
and financial fragility after a
windfall.
Chef
[applause]
Consumer
[applause]
loans, heterogeneous, interest rates and
inequality.
Professor Economy University
[applause]
[applause]
>> [applause]
>> in French.
>> [music]