The central bank of the future: opportunities and challenges
OeNB I SUERF Annual Economic Conference 2024 – VOWI 2024
10.06.2024 – 11.06.2024
▬▬ Contents of this video ▬▬▬▬▬▬▬▬▬▬
00:00:00 Academic Session 2
Defining and communicating the monetary policy reaction function in an increasingly complex environment
Market Perceptions, Monetary Policy, and Credibility
Vincenzo Cuciniello, Bank of Italy
Fed’s policy rule: A discrete-choice regime-switching approach
Andriy Sirchenko, Nyenrode Business University
Whose Inflation Rates Matter Most? A DSGE Model and Machine Learning Approach to Monetary Policy in the Euro Area
Johannes Zahner, Goethe University Frankfurt
Co-author: Daniel Stempel; Heinrich Heine University Düsseldorf
Moderator: Maria T. Valderrama, Head of Monetary Policy Section, Oesterreichische Nationalbank
01:17:04 Policy Session 2
Are central bank finances relevant for their ability to deliver price stability?
Sarah Bell, Head of Central Banking Studies, Monetary and Economic Department Bank for International Settlements
Ashok Bhatia, Special Advisor, Office of the Managing Director, International Monetary Fund
Dirk Broeders, Senior Lead Financial Risk Manager European Central Bank
Vasso Ioannidou, Professor of Finance, Bayes Business School (formerly Cass)
Moderator: Robert Holzmann, Governor, Oesterreichische Nationalbank
02:32:16 Keynote 2
Central banks between economic soft landing and growing political expectations
Francois Villeroy de Galhau (online), Governor, Banque de France
Moderator: Robert Holzmann, Governor, Oesterreichische Nationalbank
03:17:18 Policy Session 3
Monetary policy in the polycrisis new normal
Fiorella De Fiore, Research Adviser, Monetary and Economic Department Bank for International Settlements
Michal Horvath, Executive Director, Monetary Policy and Market Operations National Bank of Slovakia
Frank Smets, Adviser to the Counsel to the Executive Board at European Central Bank European Central Bank SUERF Fellow
Moderator: Ernest Gnan, Secretary General, SUERF
04:36:16 Academic Session 3
Inflation – new insights
The return of inflation and inflation risks
Juan Angel Garcia European Central Bank
Co-authors: Ricardo Gimeno, Banco de España; Piers Hinds, Haozhe Su; Michael Tretyakov; University of Nottingham
Can Supply Shocks Be Inflationary with a Flat Phillips Curve
Jean-Paul L’Huillier, Brandeis University
Co-author: Gregory Phelan, Williams College
Maximally Forward-Looking Core Inflation
Karin Klieber, Oesterreichische Nationalbank
Co-authors: Philippe Goulet Coulombe, Université du Québec à Montréal; Maximilian Göbel, Bocconi University; Christophe Barrette, Université du Québec à Montréal
Moderator: Alex Grimaud, Economist Oesterreichische Nationalbank
#oenb #suerf #vowi2024 #centralbank
Find more information on this event here: https://www.oenb.at/Termine/2024/2024-06-10-volkswirtschaftliche-tagung-der-oenb.html
so good morning let me welcome you to the second day of our annual conference um we start with an academic session so I think this is to so everybody wakes up with an academic session a very interesting papers um the title is defining and communicating the monetary policy reaction function in an increasingly complex environment and we have three uh outstanding academics three very good papers um and I think we will enjoy that um each of you has 15 minutes and I have repeated I’m very strict about it um venu the floor is yours okay thank you very much for uh I’m included my paper in uh in the program so basically this is a paper about monetary policy Rule and how this rules are perceived by financial markets and moreover we’ll focus on how this perception about the the response of the monetary policy towards inflation can affect the credibility of monetary policy so as you know uh since 2021 almost all central banks increase interest rate however uh we know that when agents are forward looking what really matters for you know assess the monetary policy transmission is not only the actual level of interest rate but also the path of interest rates simply because agents are forward looking and also they want to internalize what is the commitment of the central bank to uh long run inflation Target so in my talk today uh basically I will try to answer two questions so the first is has the financial Market perception of the ecb’s uh reaction function change over time second to what extent does the perceived monetary policy responsiveness to inflation contribute to the ecb’s credibility in order to answer this question I focus on high frequency data in order to estimate the elasticity of the expect nominal interest rate to uh inflation expectation which is captured by this parameter Pi then once I have these estimates I can also calculate if you want the state uh contingent pass through of uh short-term inflation expectation to Long Run uh inflation expectation in order to see how uh this pass through change according to the level perceived the aggressiveness of the monetary policy reaction towards um inflation just to give you a preview of uh results I find a significant sizable change in five uh after you know the the starting of the um uh hiking cycle of ECB moreover I find that higher is f the lower is the pass through from short-term inflation expectation to Long Run inflation expectation okay since this is an empirical paper I want to uh simply describe uh the data that I used so I used both Financial based and Survey based data on the Euro area since 2005 to 2024 I have daily data for short run expectation in particular uh I’m going to use for inflation expectation inflation work rates for uh expecting nominal interest rate uh I’m going to use uh uh overnight uh index was rate and for I have also Daily data for the forecast of GDP growth rate this comes from uh The Continuous consensus forecast as to Long Run expectation I’m focusing only on survey data and uh here the hor Horizon is six to 10 year forecast um of inflation GDP growth rate and three Monon uh euror EUR reor rate from consensus economics maybe I can skip this I mean this is just you know the row data that uh uh I use it uh I’m serious what I want to focus is is is this um on on on this SL is the also the motivational slide for for for for my work so if you look at the correlation between uh the um overnight index swap so the expecting nominal interest rate and expect inflation if you see the full sample here there is a positive correlation between the two right but if we concentrate on the uh sub period we can see that in the period uh between 2020 and 2021 this correlation was you know was positive but was almost flat right then when expected inflation start to you know go uh goes above you know 2% also markets start to you know to um to think I mean to I mean to to price a differ sens itivity if you want of expected nominal interest change to a per to a unit of change in expect in inflation expectation in other words you can see that the the slope of this uh of this curve has changed in particular this is much steeper and not only the slope but also the intercept has change right so this suggests that you know maybe a financial markets have perceived you know a change in a monetary policy regime in order to do this more formally so I’m starting from a very you know standard uh Tor rule which is equation one which I is the nominal interest rate then I have this parameter row which is I mean the capture the uh inertial component in the interest in in the rule then I have the longrun interest rate this is the inflation Gap this is the output Gap this is a monetary shock since I’m assuming that Financial Market perceived I mean that this is actually the the Central Bank reaction function this must hold in all Al in in Period K so I can take expectation and basically I can estimate this uh this equation and I’m going to focus on uh on one year one year ahead as a shortterm uh Horizon however I mean if I simply run a regression this equation this is a simple correlation so I need some identification in order to you know to have some result some uh causation relation between uh inflation Gap and uh and uh interest rate in order to do that I’m focusing on three-day Windows uh around the hicp inflation pressure relase and why I’m doing that because I find that in the data there is much more volatility around these dates both for inflation swap rates but also for GDP growth rate so this suggests basically that uh Financial participants are really updating their uh expectation about uh uh you know inflation and uh and GDP growth once they receive news macro macroeconomic news because on those days we have you know inflation news but we have also GDP and unemployment rate news okay so this is our I mean uh summarize the the first estimation that I run so in the First Column there is is or less uh estimation uh just on the three days window around the hicp inflation fres release on the second column I estimate the um forward looking ter rule on the whole sample and I’m using as instrument macroeconomic news just release I mean the surprises that um are generated on uh are calculated on hicp inflation FL releases and as you can see basically the two approach are quite similar so they produce more or less the the same results what I want to check uh then in column three 4 F5 is to see if the uh the coefficient uh the coefficients uh uh have change over time and here basically you can easily see that for example if I take colon four and colum 5 this is the the coefficient attached to the inflation Gap we have that the coefficient in 2 uh 2022 202 24 is at least three times as large as the coefficient in in the period 2013 2021 so this seem suggest that you know some sort of structural change or you know regime change uh uh occurred another way to visualize this this results is to run the the the same OLS uh regression but using uh a rolling Windows around hicp uh uh flesh release date and here you see uh I mean the the main results of the paper basically so that uh over the the period in which policy rates were you know close to the effective low bound and you know there were was a period also in which uh you know asset parchas were were implemented so the the inflation Gap coefficient was very close to zero when the ECB started actually to uh increase interest rate which was here in this period here so also the the perceived aggressiveness towards inflations uh changed and you see that the the change was quite suable also significantly different from uh from the previous period I want to show also another results which is related to the uh inertial coefficient because I mean the two seems to be quite symmetric so in the period in which the F50 lower bound I mean policy rate were close to the effective lower bound there were uh there was forward guidance also the stickiness of uh uh nominal interest rate was quite High it is quite intuitive right so because basically the the ECB communicate that interest rate will be you know low for a very long period but once you know the communication and also the the action of the Central Bank changed so when the start increasing interest rate also the the inertial coefficient uh change actually reduced and this I mean I argue is in line if you want with a data dependent meting by meeting approach so it seems to suggest that more volatility um uh actually uh occur uh in in in this period okay so once I have this um estimate what I can do is simply to uh calculate the um the pass through uh from shortterm to long-term inflation expectation conditional on uh on F hat in particular what what I’m doing here is to uh this is on on the left hand side is the daily change in a five year five year uh inflation swap as a short-term inflation expectation I take the one year uh uh daily change in one year inflation Swap and here for example the the beta coefficient is if you want the unconditional uh pass through of uh uh from shortterm to Long Run inflation expectation then I have this other coefficient here which is Sky which capture the interaction of my estimate coefficient times the the change in a short run inflation expectation and what I find is that this coefficient is negative so in other words what happen is that when Financial Market perceive that the fight is high also the pass through is is lower and you can see this also in in this picture here which on a horizontal axis I have the quantile of the distribution of five head and basic and on vertical axis I have the full pass through from short run to Long Run inflation expectation you can see that when we are in you know the top uh uh um button of uh the distribution the pass through is very close to zero so this seems to to suggest that when we are in this uh uh regime basically the The credibility if you want of the central bank to uh to to to Target a long run inflation expectation is much higher than in a when we are for example in a in a bottle uh bottom of uh of the distribution so summing up basically I uh show you a methods for testing the perceived shift in monetary policy rule I find that um after 20 uh 22 basically there was a a shift in monetary aggressiveness uh on monetary policy toward toward inflation and this perception also can affect the pass through of the short-term inflation expectation to Long Run inflation expectation and this evidence seems suggest that a more credible commitment also um to Long Run inflation Target occurred in uh in this period thank you thank you very [Applause] much thank you very much uh great presentation so now we have 10 minutes for questions from the public or online uhan and then jaob very nice uh Pres presentation and very nice paper I have a quick question about the regression you run if I read it correctly you had a Pi Bar right so and you were taking the difference to Pi Bar um so were you assuming that Pi Bar is equal to the Target and uh if you were to relax that assumption as for example some time varing um Target as perhaps trying to capture the idea that when inflation started to increase perhaps some Market participants were a little bit skeptical that the central bank would be able to achieve its Target is I don’t know how the the scope that you have to capture that type of uh mechanism I know you did the last regression on the long run expectation but you know you were constraining it seems in the first regression Pi Bar to be equal to 2% so I was just wondering on you know how how to think about it thank you thank you for for for for the question can I answer now so basically here in a in a baseline I’m taking all long run expectation from surveys so this is the the consensus uh uh forecast for Long Run inflation expectation but also do the exercise uh in which I impose that I mean the Pi Bar is equal to 2% but I mean doesn’t seems to to to change because these are you can think this is you know as a low mov oh sorry a low moving uh variable so in that interval that doesn’t seem to to matter uh uh a lot so yeah okay Jacob please here he was first sorry I have uh two questions um one is a small question about the coefficient of the output Gap in your uh the last period that you estimate because it seems to have the the wrong sign and it is still very significant so what’s the story there um and the second is uh and I think a bit in line perhaps with uh yesterday’s keynote speech by by uh Professor oranz um you know my perception of uh a central bank following a rule implies that uh you you have a benchmark which provides you with um you know the normal kind of policy rate and then you can deviate from it if you have a good reason but if that’s the perception of a rule right how how does your finding that the coefficients change so much over time should be interpreted my my interpretation but it’s challenging perhaps is that okay there is no rule apparently because if there’s a a rule you know these coefficients should not change okay so uh first question so I I don’t have here uh in slides the the but you’re right I mean in in the last period the the sign is negative this seems to me uh like you know you can interpret this as you know some sort of Supply shock where more relevant in in this period so because also what what what I found is that if I look at the relationship between the uh the output Gap and U and uh and inflation uh sorry and the uh nominal interest rate expectation I find a very strong negative relation also in uh you know in in raow data and this is seems to be you know that um also related to to the fact that basically forecaster you know are assuming because this is a one year one year ahead GDP forecast so this after co theyum uh there there anticipated you know a strong recovery from from from from the covid and uh this seems to be quite uh quite important in uh in the data um as your interpret so here of course I mean this is that the key assumption is that I mean this is not the actual rule this is what Financial Market or financial participant believe I mean or they perceive to be you know the the reaction function of of monetary policy and um what I’m I I I argue is that this belief can change over time so is not at this can depend on communication but also uh I I think also the the period in which we were the zero bound even if these are in terms of I mean we are talking about expectations so somehow you know that the the zero bound should be less binding if you want when we talk about expectation this seems to to matter and the idea is that I think is also very connected to the forward guidance because when you know you you provide you know an indication where the Future Path will be and will be low for long this of course we going to going to affect also um you know what will be the the future expectation about a reaction to uh to inflation so yeah okay the next question please aanas Oran very interesting work and the data and the uh U and the window you have outside the hicp announcements I have uh two questions one is uh associated with the difficulty of U doing leas squares uh regressions uh when you have an effective lower bound that we know in the case of the ACB has changed you had negative rates for a while but still for about a decade um you have the problem of the ACB might have wanted to lower the interest rate more than it could and uh the dependent variable you have cannot capture that which will lead to a bias in the estimates including the response to inflation that is that is so critical for your analysis and I wonder if you can tell whether the U disappearance of a response to inflation for a while can be attributed just to the fact that the effective lower bound was was in effect uh I wonder it’s it would be imperfect but perhaps uh if you could uh maybe you have already done it uh check the analysis with shadow shadow interest rates as has been done in some of the studies and see see if that can deal with the problem again I’m I’m not aware of a good solution but you could at least find out to what extent the coefficient you’re looking at is biased the other one that I think is is is interesting to follow it would relate to the question that was just answered in the um Post in your answer um the forward guidance period um so if you check the period from say um U second half of 2021 to second half of 2022 uh the policy rate was unchanged but uh inflation and inflation swaps were jumping through the roof can you identify if that period was a special period where you can say yes whatever the underlying policy rule that the ACB was following before and after that was a special period when something went wrong and that would be an example of deviations from uh from the perceived policy rules so I wonder if you looked at that period specifically yeah no thank you thank you for questions so starting from the second one because actually in that period uh I mean was a period in which real um uh interest rate were negative so I mean this is uh undebatable so uh even if I take you know forward looking uh uh expectation there were negative uh however what I found is that when actually you know this the ECB uh communicate and normalization we was in December 2021 something changed in in uh in a market perception because is that actually the period in which I I I I observe you know a change in uh expecting nominal interest rate but I I agree with your point so I don’t want to say that in that case I mean from in that period the um monetary policy was restricted no no not at all so was still a commodi but I see a change in um in the financial Market uh perception about what will be the future uh reaction function of of of the Central Bank and agree with you I mean the the the zero lower bound is uh clearly you know uh an issue here in terms of identification what I try to do I mean one goes back to my previous reply so usually I mean zero lower bound should be in principle less binding when when you talk about expectation no somehow the problem is that here was a very long period and that’s what forward guidance also uh comes in and um what I did is try to to estimate also with a damy so this is also another way in which uh usually people try to to tackle the issue of zero bound and this seems to to hold so basically I put a dummy in the period which the interest rate what that zero over bar I didn’t use actually the the shadow interest rate this uh this could be you know something I can do another uh check that uh um I did is this was also in in the table I split I I use a sub period so I concentrate in the period in which the the interest rate were you know not very close to the effective lower bound and um I found that more or less that the results are quite similar so that means that I mean what is really key in other words is uh it seems to be you know really a regime change when uh when when the interest rate goes to to the zero lower bound so this is really the S so when they enter in this regime when they exit and this seems to be uh really a key a key driver of uh these results I don’t have enough data actually on on previous I mean before the global financial crisis I also have some issue there for you know liquidity Premia risk premum because in that period I have inflation swap but um what I did also is try to to use um consensus forecast expectation also for for for inflation and I mean the period before the zero overbound and after zero overbound are comparable so it doesn’t seems to be what is really different was the period uh you know of unconventional monetary policy you know when uh asset partch uh asset purchases uh program were implemented for our guidance and so on so forth we kind of running out of time but Frank has a question no just to follow up I think it would be it would be good to actually do the formal test no from from your time varying results it looks like there’s the zero lower bound where you would expect it to be no close to zero and then before and after is the same reaction coefficient and so that it would be good to actually do that formal test yeah and so there’s nothing special about all of this no it’s uh when you have a lower bound then of course interest rates cannot adjust um but when you’re out of it either before or after you have a strong response and if that’s the same it suggests that also the reaction function was the same yeah thanks okay thank you thank you very much thank you so Andre so next presentation I am delighted to be here and talk about Central Bank reaction functions in front of central bankers and monetary economists one uh of the main reasons to estimate the policy rules is a search for better policy in order to improve it h we need a clear empirical identification of what is going to be improved we need a formal quantitative uh empirical edification of the current policy obviously uh the policy interest rates are also of great interest for Market participants as a curious econometrician and outsider I tried to mimic the fat policym process as close as possible in particular I use the federal funds rate target uh as a measure of monetary policy instead of the effective federal funds rate and uh I estimate the FED I use the FMC data FMC meetings data frequency and estimate the FED response to in to an information set available at each FMC scheduled meeting avoiding the distortions caused by time aggregation neither monthly nor quarterly interval matches a natural FMC decision making cycle as you know FMC meets eight times per year moreover I try to accurately synchronize the FMC decisions with the realtime data truly available to the policy makers at each policy uh meeting rather than the latest currently available versions of the revised data I address the discrete and sensed nature of policy interest rates by using um discrete Choice econometric model during the last four decades the FED has been setting the policy interest rates in discrete increments of mostly 25 basis points a quarter of a percentage point the despite the regime switching behavior of central banks the policy interest uh the policy rules uh typically are estimated by using a single equation model that does not allow the Central Bank actions to be generated in different regimes despite the wide use of regime switching approach in macro in macroecon in macro economics so I propos a model that allows for switching among three latent policy regimes let’s have a closer look at the data these are the fomc decision made at scheduled meetings since 1987 until the end of 2018 um prior to 1989 the FED changed the target rate by increments of 6.25 basis points since 1989 uh all fed decisions um all changes um were made in multiples of 25 B basis points and basically almost all fed decisions since 1989 fall into five categories 50 basis points 25 basis points cuts and Hikes and no changes so with just very few exceptions for 75 basis point changes so I classified these 250 FMC decisions into five categories large cut small cut no change small hike and and large hike notice that about 70% of FMC decisions are no change decision even before the zero law bound more than half of policy decisions were no change decisions at the scheduled FMC meetings and the proposed regime switching model is actually um parsimonious version of fin mixture of audit probit models with three latent regimes um at the bottom of this decision three you see the absorbed changes to the Target rate no change cuts and hikes in the black box there are three latent only partially absorbed decisions represented by three audit probit latent equations we have a regime decision and we have a loose we have a output outcome decision in the loose regime and outcome decisions in the tight regime uh the model allows the hikes and cuts to the rate to be to be generated by distinct processes the the cuts can be generated only in loose regime the hikes can be generated only in the tight regime but the no change decisions can be generated in all three regimes loose neutral and tight so the um in this empirical application to the fat policy rule the regime equation the regime switching decision is driven by two variables uh an indicator derived from the policy bias or balance of risk statement made at the previous FMC meeting which is equal to – one if statement was tightening zero if statement was neutral and one if and uh minus one if the statement was easing and the tbook formerly greenbook projection of the total number of new housing units started in the current quarter it’s it’s uh it’s well known as a leading indicator of economic uh strength actively monitored by by the fed the outcome equations are driven by two forward-looking failure rule variables the tbook projections of core CPA core pce uh inflation rate and output gap for the next four quarters so this is the result of the estimation for the entire period until the end of 2018 but and and the model clearly outperforms the single equation audit probit and linear or less regression with the same four variables included but let’s see how this model performs out of sample so I estimated this model for the period until the end of 2 of 2018 when Target rate rate reached zero and then I performed out of sample multi-step ahead for cast of the next 80 FMC decision during the next 10 years so these are without uh reestimating the model as you can see the average forecasting arror is about five basis points if you compare uh the forecasting performance of the single equation audit proit and linear regression you can see that linear regression and audit probit has forecasting error which which is three or four times as large as the errors of the regime switching model I also performed forecast with recursive reestimation so I added uh FMC dision one by one reestimated the model forecasted the next decision and I did this recursively and the performance of the in this case performance of the single equation models the forecasting errors are twice two times or three times as large as uh with the switching audit proit model well then um I I also Tred to to further improve the uh performance of the policy Rule and instead of um projection of the course CPI inflation I used um term spread term spread between the one-year treasury rate and effective federal funds rate which is commonly seen as a low dimensional market-based precursor of of future inflation and economic activity and um the performance so this is the estimation of the model when we have term spread instead of inflation rate and let’s see and let’s see how it Compares with the model with with PC pce inflation as you see in Sample fit is much better for the model with storm spread likelihood and information criteria are sign significantly better and the um forecasting error is smaller by 30% this is the in Sample fit again until the end of 2000 of 2008 and now let’s see what happens if we do out of sample for cast for the next 10 years then performance of with with storm spread the forecasting error is 20% lower than with inflation rate and if we do out sample forecast with request of reestimation the forecasting error with term spread is 10% less so we actually able to um forast the U forecast the FMC decision during the zero lower bound and after the zero low bound just making uh average Arrow of three basis points um okay uh the next step is going to be since we know that um this this rule is explains how the fat reacts to information set but it’s not operational it’s not available for the public because the the tealbook projections are available with only five LS so to make it operational for the public The Next Step would be to estimate uh the policy rules for until now using the publicly available data uh forecast made by made by um s SPF forecast realtime data on the housing starts um and see how the the how the polary rule with regime switching performs compared to the conventional um single equation OLS and single equation discrete Choice models thank you very much for your time and attention [Applause] so we have time for questions who wants to start tasas here please thank you aanas Oran very nice work and clearly since changes are 25 basis points are multiples of that then discrete policy approach uh should work better as you show um a couple of remarks and a question so marks is that pre1 1989 U unfortunately it’s not quite the same regime because um the um um the FMC was operating with a with a range uh and sometimes they would just adjust the upper or the lower part of the range this is why you get this six and six and a quarter basis points changes so maybe it may be more useful to start the estimation after the Thanksgiving turkey episode of 1989 which for fored the the committee to move to 25 basis points clearly communicated policy changes um the question I have is um when you estimate uh this process have you tried also the growth Gap instead of the uh output Gap or alternatively um the projected change in the unemployment rate uh instead of the unemployment Gap so this is the difference between U uh grow growth versus levels uh in in policy rules with more parameters usually both matter and it’s it would be useful to see if within discrete Choice model you can shed some light about the relative importance of the level Gap versus the growth Gap this is one thing um and um the um the other thing is that it would clearly be fascinating if you if you um uh did as the next step what you just suggested taking the quarterly SPF projections that are available for the Last 5 Years it would be fascinating to see uh If the Fed has been operating with the same policy rule or uh we have this period of a couple of years when we have these massive deviations that were quite quite unusual from historical uh uh data we have because the SPF goes back to 1968 much like the greenbook projections go back 1968 we know that the the SPF uh median projections are very similar to the greenbook projections so it would not introduce a whole lot of error uh if you used the SPF projections to do an analysis for the last five years yeah thank you um by the way I’m very delighted to see you because when I was a PC student at the FED uh more than 20 years ago I started doing this research on U after reading your paper on policy rules using real data and what you said yeah I I just I just showed you the just some results um based on the fomc data uh yes I’m working now on um fed policy rules using publicly available data first of all SPF I am aware of this data and yes indeed indeed the the preliminary results that that I got uh was very was very good and uh are you talking about a growth Gap uh it’s a change in the growth rate and change of the GDP grow it’s a a projected growth rate of GDP compared to potential output growth okay thank you yeah I will try that and definitely I I’m looking at all all available publicly available data uh in especially in the forward looking context first of all SPF data is yeah I I’m I’m applying it so yeah I hope to to have the the next results for the International Association of for the next meeting in Talon I I will present the results hopefully with publicly available data till the end of till till till 2024 yeah uh Katherine please thank you the results amazing but I was wondering why don’t present you have a lot of observations that so I think any model will conate oned no interest rate change which is quite fine for the period of the so I was wondering why not showing the number of corrected at well I have just 15 minutes I showed you the summary of the um of the results the overall accuracy right the percentage of correct predictions so I I could show you also the the scatter plot where you have predicted predicted uh decisions and actual decisions across the time but yeah and but you you are right that prediction um prediction for the zero lower bound um the regime switch and model immediately predicts zeros so the neutral regime which one of the three regimes of the zero of the regime sing model immediately um accommodates these no change decisions when you do ol regression it takes some time until the OLS or single AIT proit Laars that actually we absor 000000 and they just this uh single equation model they after some period of time they just switch to predicting zeros forever because it it makes better likelihood uh so but but the regime s model predicts zero immediately and that’s why the um overall prediction during the zero lower bound is is just uh uncompatible after the zero low bound the difference between regime switching and single equation model is not so huge but still better but during zero lower bound yeah it’s uh the switching model basically I think it it the prediction accuracy is 90% during the zero low bound which is not a big deal of course it was easy to predict zero at that time okay okay uh Stan C again one short question is uh can you be a little bit more specific on the way you estimated the latent variable that that governs the reim switching is it just is it a cman filter or pure latent variable or or how do you estimate it this model is estimated just by maximum likelihood H the extended version of the model where you have Auto regression when you have when you have lacked latent when you have lacked dependent variable cannot be stimated by maximum likelihood then I use I then I estimated it with gib gib sampler using basan methods but this model is just it gives you it it’s estimated by full full maximum likelihood it gives you the consistent estimates not only of the probabilities of absorbed changes but also it gives you consistent estimates of the probabilities of all three regimes okay if there are no other questions are there any online questions no okay okay so thank you [Applause] so I should just press next y wonderful okay hello and welcome thank you very much for including me in this wonderful conference this is a paper with my colleague Daniel stemper from the University in dorf and I was just told that I should keep the presentation as less technical as possible I will try to do my best um but what we do in this paper here is we combine a DSG model approach and a machine learning approach to answer the question um whose inflation rate in the Euro area matters most for um a systematic policy rule by the ECB okay let me give you um and I probably not have to tell you that um inflation rates matter quite dramatically in the Euro area whether that’s that’s just not now the case and here you see that in this on the right uh on the yes on the right side uh you see basically 20 shades of blue where um there are some countries having quite High inflation rates well right now uh for instance Croatia and other countries with rather low inflation rates for instance Italy right now and it kind of raises the question um how a monetary policy maker would react when there are both at the same time deflationary and inflationary presses um but the only instrument available is conducted uniformly across all its members and I this might not be super relevant when they are uh when these differences are quite random however once um we have countries that structurally deviate from the Euro area um this might become a policy issue that’s where the paper starts to be and and so in the first part of the paper what we do is we identify two country groups that structurally differ in their inflation volatility and kind of show that in in this graph here that we have a group of count so in the graph you see on the left hand side the average inflation rate of um country me of of member countries in times of high inflation rate so whenever the inflation rate in the Euro area is above 2% and you see the opposite on the left side and and you see that there is a group of countries that whenever the inflation rate is above 2% so it’s we in a high inflation rate these countries are above that average and whenever the inflation rate is below 2% so we are on a low inflation rate um these countries are below the average so on average well they have higher volatilities and and well in particular that’s Greece Ireland Italy Portugal and Spain and then we have the opposite we have low volatility countries so whenever the inflation rate is above the average these are below and whenever the inflation rate is below the average these countries are above and particular that’s Austria Germany and the Netherlands and so for the rest of this talk I will focus only on these eight countries and I will refer to the sub group um of high volatility and low volatility now we could ask well whose inflation rate matters more those of high volatility or low volatility and I mean a standard way to assess that would be an interest rate uh systematic interest rate rule uh that’s a very simple uh rule in in fact where we only include um deviation from uh two deviation of the inflation rate from 2% and if we just weigh both high and low volatility equal um we get this ah wonderful we get um this this well this ol regression in in the middle and and I know that you are drawn to look at the inflation response coefficient which is absolutely fine but I want to highlight or I want to put your focus on the R square which remember tells you how much variance um this regression actually explains which 43% so quite a good chunk now the interesting thing is once we put more weight on high volatility countries so we weigh their inflation rate disproportionally strong this coefficient actually starts to increase now we can now explain I think it’s around 15 yes 15 percentage points more of the underlying variation now if we do the opposite exercise we now focus on put disproportionate weight on low volatility countries the coefficient shrinks it doesn’t shrink to zero meaning that there is still uh a required a good amount of variation explained however it’s definitely it’s sub substantially and significantly lower than an equal weight now the question we are asking therefore is which weight is now accurately describing the historical emu monetary policy Rule and given that this weight on high and low volatility countries potentially not not fixed but rather flexible um most standard empirical estimation approaches uh will be quite it will be quite difficult to to to catch that weight and so we came up with a we called it a data driven solution because it sounds nice and fancy but it’s simply the the following we build a standard new Canan model so a standard dsge model with well we have later three monetary policy regimes we have a regime that focuses on high volatility countries regime that has puts equal weight on both country blocks and the regime that focuses on low volatility countries and we simply generate data from these three regimes and what we then do is we train a machine learning model we use that generated data and randomly sample from each these each of these regimes and whenever and we ask machine learning model well which of those three regimes do you think the current um the current data actually stems from and and we can get a model to identify correctly 97% of of the time and then we are really happy and instead of using the synthetic data we then actually use actual emu data and ask it well what do you think over the past 20 years in each quarter what do you think is the current regime we are actually in and we have two two results first we find an a biased weight on high volatility countries and then we show um that the ecbs reacts more strongly to countries where the inflation rate de deviates more strongly from their long-term Trend okay so this is the only slide on the methodology so um if yeah it’s basically doing what I just told you so we have a simple currency Union two countries household sector firm sector you see no belts and whistles here we have a monetary policy rule looks um it looks very standard straightforward with these three uh Central Bank regimes we calibrate both um uh we calibrate the countries to correspond to these eight countries that I showed you earlier and then we simulate 10,000 periods for each of these three regimes we get a bunch of macro data so information on each consumption information on the inflation rates and so on we split the data into at20 and to an 020 ratio we train the model on 80% of the data we do out of sample forecasting on the remaining 20% we evaluate and train everything in our case a neural network performs best um and then we use that to classify emu data that’s it okay now so this is already these are the the the initial results you will see that we will do some more stuff afterwards but this is what you Pro uh this is I think the the probably the most interesting and relevant slide okay so what do we find well first ignore the foreground ignore the lines in the foreground only look at the background so what we did is we shaded the background on the weight that the neural network predicted so whenever you see a dark shade so in the beginning for instance this is when the neural network told us well I think that given the current data um the focus of the ECB was NE had a neutral stance so it put equal weight on both country blocks if you look uh Whenever there is this medium Shades this is when our Neal Network predicted that the focus was on low volatility on a low volatility country block and for the remaining time the focus was or the weight that the neural network assigns was on high volatility countries and so you can immediately see that there seems to be a disproportionate emphas emphasis on high volatility countries inflation rate is roughly 80% of the time now um when when when when you spend as much time as as we did so clearly not possible in a 15-minute presentation and and you start looking at each of these IND individual deviations you might start to see a pattern there and I will just try to to emphasize that so let’s for instance look at this regime switch uh towards low volatility countries and that kind of uh that kind of starts when the deviation when the straight line is below the dotted line now the straight line is the inflation rate of or the deviation from a long-term Trend the inflation rate of the low volatility countries and the dotted line is the same for the high volatility countries and you see that the regime switch appears when there is a greater deviation for the low volatility countries and the low volatility countries got the the weight of the DCP oh god um so the and that reverse roughly in 20 when the deviation shifts to the high volatility countries so and that kind of stays the same for the next seven or eight years and so we find anecdotal evidence that the ECB is simply reacting more strongly to Greater deviation from inflation rate um which is a potential explanation for one but very importantly it’s a potential confounder in the in the in the analysis that I just conducted and so to okay so so just to clarify ideas here um if the ECB would actually react to individual deviations rather than aggregate ones we would have a slightly more complicated reaction function and ignore the math here uh but what that reaction function would tell you is that whenever a c that there is a true weight on both country blocks and whenever there is a deviation from the average um so meaning that one country has a greater say higher inflation rate than than the other country block um then there is the the EB shift the weight towards uh this this country block meaning that well there is a true weight but that true weight is the observed weight that we can observe is biased by um this deviation behavior and so the next thing we do is basically we refine our whole process so that we get weights in continuous time so we want to have this large Omega continuous time so we can estimate the actual uh we can estimate this this latter this this L equation and we can get the true weight that the ECB play on each of those two country blocks and estimate that and let me ignore how we do that um uh but what we get once we have the the the weight in a continuous in continuous time is we get this density for for the weight and what you see is so a weight of one would mean a weight on high volatility countries weight of zero would mean a weight on low volatility countries and you see that on average the weight is the observed weight is I think 67 so 60 67% however remember that this weight is biased and so what we do then is we run a simp or less regression where we have the deviations on the right hand side and we have these observed weights on the left hand side and this gives us I mean this equation is basically the same as the one I just highlighted earlier um and the nice thing is that now we can interpret the constant as a non-changing weight over time and we can have this second part of this of this regression as the uh uh as as this change in weight but not due to an inherent uh favorability towards a certain uh toward certain countries but simply due to the fact that these countries deviate from their long-term Trend so what would be our expectations well we would have prior expectations that the ECB puts a weight of 0. five on both country blocks um and that it indeed reacts at least given our anecdotal evidence earlier indeed reacts more stronger to Greater deviations and so we can run that regression it’s it’s quite straightforward what we find is yes indeed let me start with the constant remember we can interpret the constant as the true weight that the ECB puts in this case on the high volatility countries it’s 62 and remember that earlier I showed you the observed weate was 65 so around five percentage points can be explained by this uh change in loss function but basically by um this this deviation Behavior so yes in okay indeed we find the biased weight um by about 12 percentage points but we also find this deviation from uh inflation rate uh um behavior in the last column I just showed show you that it doesn’t matter for what we control for we actually control for a bunch of more stuff doesn’t uh our results remain rather rather the same okay so here is a simple conclusion we try to figure out whose inflation rate matters most for ECB monetary policy we combine a DSG model as a data generating process and machine learning model um as the data receiving process and then we use the stuff that machine learning learned uh and feeded historical emu data between 2004 and 2021 we find the disproportional emphasis on high volatility countries but we also find that the ECB simply reacts more stronger to countries when their inflation rate deviation from the L from the long-term trend is more substantial with that thank you very much for your [Applause] attention Okay so we have time for questions y jaob danan um there is an old literature on on this issue as you probably know um and the official ECB policy of course is that um what happens with inflation in individual member states doesn’t matter what the ECB uh supposedly is doing is focus on the Euro area inflation right and and your result suggest well they don’t um at least you know um there is a um some bias in that sense right um initially you started with several models showing um weights of High versus low inflation uh and what I was missing there was what I would consider The Benchmark to compare your results with namely uh a tailor rule model where you take the Euro area inflation uh instead of you know the high versus low volatility weights yes so so how would that Mark model I’m I’m sure you did it yeah how does that model compare to your uh other models with high versus low volatility weights okay um say I’m not entirely sure to which because there are several literatures that yeah but let me let me answer the tailor rule question first um yes I mean we run a tailor rule uh we run different tailor rules uh um this one that I showed you is merely for dactic reasons so I would not use that as actual evidence for for differences in weights for instance as you correctly I mean I think I wrongly highlighted as an emu tailor but anyway um so you’re you’re correct I mean there is lots of stuff missing so an output Gap and and and inertia and so things we use that doesn’t change the results it just makes the presentation a little bit more demanding um on why we think there might be different weights now I have to say there is a carard in there um we run that only on eight countries um because well then we can shift the weight uh we can’t run that on 20 countries because we would have to I mean for instance Croatia would have a really weird weight um of zero and then just appearing in in the last in the last moment um which would I mean that would clearly bias bias our results because then we would have this yeah so we we run that only for these eight countries countries and there we have tailor rule we have a tailor rule there we didn’t run that for all 20 member countries and the average there because we have at least we haven’t properly figured out how we can then get the the the coefficient on the explained variation uh proper interpretation of of of that value but I would not place too much value uh too much emphasis on on on this this this first tailor tailor Rish regression on the on the literature so we run a hell lot of of robustness tests both from political economy explanations to to others um expectations and and and what have you and and we found that whatever we throw into this final regression the both the I mean the two coefficients that are really interesting for us the constant and the inflation respond they remain constant so at least we couldn’t figure out another way to explain explain the results you Ed thank you very much for this extremely interesting presentation I’m trying to interpret it and assess it from a polic perspective and I’m really interested in your um in your view on this so does your paper suggest that the ECB has a bias towards uh what is going on in Mediterranean countries if I may say so or alternatively might such nonlinear reaction function as I interpreted be welfare optimal if also the cost of deviation from the inflation Target is nonlinear yes to yes to the second part um so neither Daniel nor me are legal experts as far as so we we we try to figure out as far as we can tell that L function that I showed you earlier is consistent with the ECB mandate um the it so that so first thatuh so the deviation part I think is is not generally an issue um whether the because we can’t explain the whole bias with that so this this bias remains and and there might just be stuff that we haven’t controlled for um that but I’m I’m very worry to uh uh to to say that the ECB is bias towards Mediterranean countries um I can only tell you that from our regression we can’t remove the last 12 percentage points so this is and but potentially I mean there might just be take take it as you as you want to intere okay any other questions Alex hi uh I understand that you had to brush over the technical side of your paper but what I still haven’t figure out is why do you use artificial data to train your machine learning and why don’t you just run just a basic ban estimations and give you the change in regimes and the change in bias of the ECB see did would work for the first part of the paper but as far as I understand it would work very not as well on on the second part when you get the continuous weights um so yeah that’s that’s kind of the answer so the first part yes you could just you could use regime switching I mean that’s exactly yeah um on on the continuous weight I think that would introduce quite quite some someing at least on my part okay so that’s why you you have to train on this artificial time series basically yeah okay thanks okay there’s a question in the back Mainer University of cards I a very simple question so there is this bias towards higher volatility countries is this good or bad what are the consequences of this bias so we remain agnostic in the paper um and my answer will be as agnostic I as I can phrase it um actually we we yeah so um I I don’t know um as I said I mean we tried a we honestly tried a lot to to reduce the buys as much as possible um we we then went to literature of poit political econ literature um that uh well suggests that say composition in in boards matter doesn’t matter for us so it doesn’t seem to be a bias that is driven by um in this case the over representation of high volatility countries so yeah we’re trying still to figure out how to explain the last 12 percentage points and and I’m happy if anyone could can can up with a better answer than than us because I mean um I don’t truly believe I mean I as I said just the result but I don’t think there is this inherent true heterogeneous weight on on on uh on countries in in in the ECB Council but maybe um you can ask around and you find different answers on that I know I was hoping Frank says something no I I was just wondering no I mean one interpretation is that the information content of the sort of the more volatile countries is is for aggregate inflation is higher than not the information content of in inflation rates and and and maybe other variables for uh for low volatility countries I was wondering does your methodology allow for such an interpretation uh or or or not no I mean that’s a that’s a good that that’s a good very good question um I don’t think so we can only I mean we can only observe the reaction um and this at least I on the top of my mind I I don’t I kind of kind of canot come up with a empirical test that would test uh this this prediction I I wrote it down and if we can once we can come up with an idea uh uh it will it will definitely be part of the paper because I like that explanation very much is there an online question have an online question uh from Ariel schert any suggestions what drives the ECB reaction EG role of the country of the president and or chief Economist yes so we tried that so initially actually this paper was an um policy uh so it it it went into the direction of um whether the the distribution within the Council made a difference and and we we found that it does it does so long until we add our now the explanation that I showed you and then everything else becomes insignificant so we checked whether the distribution matters we checked whether um the I mean the precedency we had basically uh there’s very little variation in in precedency of course um we checked on individual roles within the ECB Council and we couldn’t find an explanation for that kind of behavior that that we do so we tested as far as I can tell on these roughly 200 um observations h on these 80 observations as much as we can and and what really stuck is the final explanation that we could come up um but yeah so we we tested that definitely thank you if there is no other question that Andis can ask again I don’t know you can hear I can hear you and I would just I can repeat it yeah okay my question is quite simple to what extent might your result have been driven by the really prolonged period of post GFC sovereign debt recession uh period which affected Southern countries much more and during a very protracted protracted period also with low inflation rates so that the ECB actually had to react because of the consistent undershooting against the inflation Target um so Matias kand asked a very similar question just earlier this week and H not ear last week and and I wanted to check that over the weekend and uh I didn’t find time so we will definitely do basically we can just train models separately over both periods and then just see whether that changed the results um I mean we I mean what we definitely do is we use um I mean the Shad rates to emulate um the uh the effective or monetary policy at the affect the flow bound but right now I mean we just Bunch the whole periods together so I can’t tell you whether there is heterogeneous behavior um across these two but now that I and more that I talk about it I mean what we have is this extreme weight during the effective low bound period uh so that might actually go into direction that that I think you’re you’re thinking where um part of that weight is basically it might be driven by uh the confounding it might found basically by uh the the post um financial crisis period but yeah we will definitely test that that’s a I I really like the idea thank you okay and there’s one last question from the so actually it’s quite related to to the previous one so have you tried to see if I mean the the right tail or the left tail of distribution matter for for your results I mean maybe the the reaction could be asymmetric if you know the volatility stems from more from the right than on the left I think is Rel to the previous question that’s a good question no we have not tested that but that I mean that’s something that we can do quite quite quite straightforward we have the distribution yeah um so yeah this is I mean yeah definitely we’ll do that thank you okay oh there’s one question the back juel Garcia from the ACB uh I just wanted to point out I mean on your interpretation I mean it’s very difficult to talk I mean my position is at the ECB so I’m a bit biased against this argument of the bias but there is something I mean you can check because um what is important is the flow of information before making the monetary policy decision so along the lines of what FR Vincenzo was doing I mean you have access to the flash estimates for certain countries ahead of the Euro Area flash estimate and there you can see that for example Italian and Spanish CPI surprices I mean sometimes reveal a lot of information about the in information about what Euro areawide inflation I mean could be so this varies a lot from time to time but you can look at these surprises and then see if your periods of different reactions only coincide I mean with extraordinary surprises I mean for those countries that is a I don’t have a good answer on that because we will definitely I mean this is something I will definitely test so we include expectations in one regression doesn’t matter there at least expectations um and what we then have is I mean realized versus expected which I think goes into that direction but yes I would definitely test that explanation um I will come to you because I’m not sure where I get these flashh estimates thank you okay so thank you very [Applause] much some people are still moving in but I think we can start welcome back uh to a very challenging session dealing with the topic uh we didn’t know it existed uh one and a half years ago and it offers me the possibility to make a remark uh as you know everywhere we go we have a lot of risk management if there’s a risk board here risk board there this and that uh they warn us of many risks but I’m not aware that anybody of these risks or whatevers want of the risk that if we change monetary policy from this accommodating policy to what’s a policy of increasing interest rate we would end up with a mismatch of our asset liability side and this is something how does not to explore why this is happening we don’t do that but what challenges evades us so welcome to a session in which uh we deal with the question whether Banks SDR still can deliver on its mandator and uh we’ll we’ll work through this question in uh question serious the first series is dealing with some of the more what is it Plumbing part of it but it has important consequences for M policy the second part is then dealing with the question does it matter that we make deficits uh this is more the more intuitive political question which is even harder to answer at least from what I know there so uh let me welcome this panel what we have here and uh it is also in this order I will then ask the questions to start out with we have here sarel is’s the head of the Central Banking status at the bank for international settlement and in this role she is also the head of the Secretariat of Central Bank governance group a form that exchanges of views among Central Bank Governors matters relating the design operation of institution so one of the experts we should be able to answer the question to right so the second we’ll ask is asop here his special advisor in the office of management uh uh director of the IMF and uh the conference he tends in his capacity course of an IMF paper last year on the profit and loss of the US system and its top five National Banks and uh he returned after some STS in in in Europe uh where he represented the was the director of them F office in Brussels and Paris then we move on uh woman men woman with Vado she’s an economist specializing in the areas of banking corporate finance and M economics she’s at currently also professor of Finance the Bas business school and the C CPR research fellow previously she was a professor of Finance at Lanas University and the still B welcome as well and last but not least uh we give uh the BRS to say the opportunity to wrap up everything at the end of and say the contrary what everything else said at news he’s senior Financial Risk manager at the financial Market division of the N bank and the full professor of fin School of Business economics at Master University so you see a very high level panel to ask to answer which question let me start out with my first question which has three interrelated elements on the mechanics of Mone creation So currently how to say we have for some Central Bank in the Euro area making losses because of the monetary asset portfolio we know this and less interest on the remunerations of the reserves this poses of course a negative track on the capacity uh and might for some like the unb Le to negative Capital now a negative capital is a position on the asset side now will this stampen the reduction of our liquidity on the liability side there one question the second one’s a little bit of a plumbing uh over in the future the effective low about becomes binding again and if central banks have to intervene again uh I say what what what does this mean does this mean we’ll end up in continuous deficit there or is the possibility we’ll become positive and then end up at the end of the day with a positive Equity or will we end up in a negative equity for the time being and the last part of it the last question is and not everybody needs to answer all the question for this reason I read them out here I to say if and as central banks exposed to uh a permanent and high deficits uh how are they’re doing it you know like the austan bank so if you’re deficit so we need to pay for it what are we doing well how to say we use the anra framework to create money but if we create money do say it means that we Finance our negative equity Capital with money creation if this happens in many all countries what does it mean then with regard to inflation are possibilities that to stabilize what we do on the one hand side with other measures there so three questions about some of the planning of Economics there without going to the detail and if out to sarra if you want to start out thank you very much and thank you for inviting me to be here really appreciate it um so I should add the caveat of course that I don’t speak on behalf of the bi yes um but I will share a more generalized International perspective than maybe a more European specific one that some of the other panelists might offer so uh just to briefly cover the first part of the question on whether it matters if the pace of reduction in excess reserves is dampened by negative capital I think the key question really to consider is whether that would affect a central bank’s ability to implement monetary policy so you know we know that there’s several reasons that have been articulated by policy makers why they might want to keep a buffer of excess reserves before reaching this inflection point where the supply of excess reserves crosses the Steep part of the demand curve um but there’s a lot a lot we don’t know about where that might happen uh partly because of our blurred understanding of banks underlying demand for liquidity or because of uneven distribution of reserves in the system so among these kind of reasons I think it’s important to to go carefully or as my boss had bis likes to say go slow because we’re in a hurry um and I think of of these reasons to be to be careful negative Capital um you know impacting the pace of reduction of excess reserves is probably not the biggest concern that we have or or the biggest thing to be to be cautious around so moving on to the second part of the question I think it’s helpful to kind of twist the question a little bit and ask if we are in a spiral how do we know that we’re in one and what what does that look like and I think one way to think about this is to differentiate the impact of the the the covid response um and that episode of asset purchases from prior episodes of asset purchases and to consider the counterfactual of of of this condition of if central banks have to intervene with APS regularly what does that look like um so first you know differentiating the covid response from the prior episodes you know I think if if we had if Co had never happened for example um would central banks be facing these losses and would we be asking about a spiral I think probably not so probably this situation we’re in reflects the particular circumstances of the covid response more than APS in a general sense um you know or alternatively as as athanasio discussed yesterday if there had been earlier tightening perhaps we wouldn’t have seen inflation go as high we wouldn’t have been facing losses so there’s lots of counterfactuals to consider I think in terms of where we’re at now um and what it means for the future so where what would it look like if if we get into a situation where central banks do need to uh make asset purchases more regularly I mean I can think of maybe two types of scenarios so one is that crises become more frequent and crisis response become more frequent maybe that’s for GE political reasons or maybe there’s another pandemic something like that I think that kind of scenario analysis probably goes beyond the scope of this panel um but I I will say I think if that’s the world we’re facing we want central banks to be able to take on the risk that no one else is taking and we want them to do that with this fiscal backing um which has which has come up throughout uh the sessions yesterday and today so so then another situation another scenario is if there’s a need for easing at the at the effective lower bound um and so then I think the question is whether or not central banks are sort of stuck in a hole because of uh the covid response that they can’t get out of um and and so even if future purchase programs are smaller or smaller scale or with less severe circumstances but I don’t think this is the case because I think for most of these and and here we’re talking about Advanced economy central banks that are facing negative equity because of asset purchases they will eventually be able to earn their way out and Ashok has done a lot of scenario analysis for Europe in particular on this um but there are some caveats around that there are some that may need to to receive some Capital injections and then I think there’s a question of long the long-term financing model for central banks which brings me I think to the to the third part of the question around the Persistence of of negative equity um which at its heart I think is p it it’s tied very much to expectations of future Central Bank seniorage income and Central Bank structural profitability um so you know as we know as long as the net present value of future revenues uh which is mainly from seniorage are large enough to compensate for current and projected losses then then it’s fine um and so for many central banks in the US in the Euro area I think cash usage is is is strong and seniorage income streams are not a not a near-term concern um but there are other central banks for whom seniorage is in Decline as cast usage is is declining quite dramatically and I think those central banks will have more difficulty and and might need to evaluate alternative income sources now and in the medium term that could be cbdc as as was discussed yesterday it could be fees um but in any case this is something that probably all central banks need to be thinking about you know either now or or or in the longer term um so I’ll I’ll stop there and others have go thanks a lot thank you how to say well defended uh I’m sure how to say there may be different views in it do you have a different view or do agree what has been said broadly I agree I feel this defense will probably continue through the panel um but as the governor kindly noted I’m here mostly because I co-author a paper Last Summer the title of which was raising rates with a large balance sheet the eurosystem’s net income and it’s fiscal implications so I’ll come back to it in a minute but really the key thing there is the large balance sheet you know I mean this is not the first time we’ve seen rate hikes yes it was the sharpest series of hikes in the history of the eurosystem but uh it’s the large balance sheet that is putting us in territory we haven’t been in before um so that paper we found that you know I’ll focus on the eurosystem we found that the eurosystem is not on some sort of path of exploding losses uh that would cause alarm um yes the losses are large very significant um we published in July we closed our calculations in March last year so yes things have changed since then but based on the assumptions we laid out in the paper then um we estimated cumulative losses for the Euros system in the 2 years 2023 24 at uh about Euro U 55 billion euros about something like half a percent of Euro area GDP so you know a very material amount people have compared it to the size of the EU budget and so on but importantly we find that uh the Euros system as a whole uh returns to profit around 2025 and according to our projection and we took a 10year horizon um the losses would be fully recouped by around 2027 um I’m completely open to the suggestion that we underestimated the losses we indeed there is a case for rerunning um we many moving Parts you know we assumed uh no reinvest uh full reinvestment of pep open-endedly because that was the announced policy at the time um we assumed a peak uh deposit facility rate of 3.5 by September it was at four um so yes uh the numbers may be worse than what we have in the paper but that was the easy part the easy part is you know you estimate roughly what the interest um outflow will be on the on the trillions in reserves that were created you make some estimates of what you know the what you get by way of a churn through uh asset reinvestment and you know organic growth of uh currency and circulation and and you come up with numbers for the Euros system as a whole the difficult part for us was to then try to divide this uh to have a look specifically at the top five central bank national central banks of the Euro area um because that was the journey that really took us to understand how the eurosystem divides things up um and the rules are in the public domain but they are not sort of straightforwardly explained anywhere so um we uh we did do that I I lots of interesting little details discovered along the way so for example when the banad Italia buys an Italian bond from a seller in Germany it is creating base money in Germany um and this then it must sort of Rec um compensate the bundus bank by paying on its Target to balance net liab just an example of the sort of things the moving Parts within the eurosystem Machinery so we found very different results for different NCBS um the bundes bank uh had the weakest results of the top five for us um with cumulative losses peaking at something like 1.2% of German GDP in 2025 that’s an amount that exceeds its General provisions and capital and reserves so indeed negative capital though not negative equity if you also consider the revaluation reserves but you know we can debate whether or not negative Capital matters but what is quite clear is the bundas bank may be looking at a decade or or or more of zero dividends to the German government and and that’s and that’s real um I shouldn’t go on too long I think in the second part we’ll discuss whether you know we’ll sort of zoom in more on our policy recommendations um but I would just say fundamental to all of this is how large a balance sheet does one want to run and that in turn ties into things like the operational framework does one want to be in a floor as seems to be the consensus nowadays or uh to venture back into a corridor system um whether one wants to remunerate reserve requirements whether one wants to raise reserve requirements um faster C um the key to us is that you know this should not allow uh this should not muddy the pursuit of the primary objective you know and just a last word to tea up for round two uh Patrick honahan said memorably last year economists are unanimous that Central Bank losses don’t matter practitioners are unanimous that they do or as my friend David Marsh said last year this is an awful lot of work for something that doesn’t matter at all so we’ll come back thanks a lot that’s a very how to say pertinent remarks at the end but for my get so far well what we had before was an exception here we don’t get Co all the time so this created lots of it but once we get rid of it the future looks bright and uh what you had in your estimates there yes uh there some problems ah had but they are manageable so this is what they get here now moving here from now to a little bit to academic side there now coming from the banking system and also what I didn’t mention is and at uh advisory she’s a member also of the advisory scientific Committee of the esap board which I attacked indirectly there so it’s my attack Justified or did you always know about it and if was this on the radar screen and so what what is your take on this okay uh thank you uh well my expertise in this uh come mostly from some work I have been doing with coauthors in the past years this has come out uh has been published last year on on central banks um accounting uh profitability or or losses and and basically we we were worried a few years back when we were seeing the size of bank balance she increasing and its composition changing to have more uh more risk in it um than we were traditionally um used to and and basically we we try to go into the data and try to tease out if we could see uh whether that that could potentially expose central banks to to uh uh problems in the future if you had some sort of supply side shock as it did materialize uh later that will generate losses we were trying to understand from the data are central banks AAR to losses and does that matter for for uh for the conduct of monetary policy and the effectiveness of monetary policy so in my in my understanding is that in theory it shouldn’t matter right so the accounting profits and and profitability shouldn’t matter so I fully I take from where you you and that um in in principle it shouldn’t matter uh but in practice there may be frictions in place which relates to the followup questions which whether that be the perceptions of the public whether will be political pressure whatever uh that could be sort of what we call fractions that frictions that may create a preference for a positive uh for a profit rather than a loss if such preference exists um then if you put it in a standard model that would lead to a deviation in policy and so what we have t have done in in this uh uh paper is we collected data from many central banks around the world over a long period of time to try to tell from the data uh whether there is such a preference for or if you want there is an aversion towards reporting an accounting loss and um and if that somehow matters for policies so um if one looks at the data it’s very clear that there is an aversion towards losses so central banks do not like to report losses if you if you if you create a um the distribution of Central Bank profits around the world you can see a very strong um discontinuity at the zero threshold so and that discontinuity varies systematically with both incentives and ability to manage reported earnings now more importantly the last part of the paper is looking into okay if and I can go later on on what factors make correlate with this aversion towards accounting losses and we also look into the negative equity um but very few central banks around the world up to the sample period had negative equity so there was not enough power to to to get any significant result there um but when it comes to policy what we find is that uh the tendency to to have a preference for profits as opposed to a loss is not independent of of monetary policy or policy outcomes the results we find are are not suggesting that necessarily um is causal like that changing uh your um uh changing the policy could be the other way around uh and so let me explain that so um if if a central bank um has on on its balance sheet assets that could potentially if interest rates increase uh lead to losses and and the central bank for whatever behavioral you know whatever frictions is facing does not want the losses it may be slow to raise in interest rates and that could lead to an inflationary pressure but the other way around could also happen that the Central Bank who um maybe their inflation rates are are are higher um they will not want at that moment to also report uh losses because the pressure uh will be stronger so in the data what we find is central banks there is a discontinuity in inflation outcomes around around zero so central banks were in the profiter to have systematically higher inflation rates um and that is on average or relative to the stated Target or relative to the IMF forast projections uh of inflation rate in the same year and we also find using a a tailor rule that uh central banks in the small profit region are systematically um maintaining lower uh interest rates than what um what the tailor rule will predict so this is the evidence but the evidence can go both ways the interpretation doesn’t necessarily mean that they change policy could be they are more concerned about uh um accounting um uh losses when some they are they are further away from from their target I could go on later on on which frictions may be creating uh this okay thanks a lot D over to you your risk manager so uh one of the risk to manage now is your post class how to say said that the estimates of of the IMF is much too low because they ignore the number of things they to share this you uh thank you governor and thank you for the invitation uh let me give you my personal views not of the Euros system uh indeed so my perspective is on risk management uh quite a few papers out there on internet that you can find to read my personal views I think there’s also a short sare of policy paper on that um before I answer your questions so your question uh so the gist of my view is centr man capital is public money that can also be used for hospitals and Roads so it should be used in a very good way for society and so negative Capital also means that’s a public deficit so it’s really important for society to manage this and also to manage the risks around central man Capital now there’s always this Mystique has so a central bank cannot technically default at least as long as it has the right to issue a legal tender in its own currency and but my risk management perspective is that central banks do need Capital to absorb risks so and we see increasingly That central banks take upon risks in times of Crisis um on their balance sheets and so what I or what I write with colleagues in these papers that from a risk management perspective you should anticipate or you could anticipate that your mandate leads to Future risk-taking yeah so not all risks are already there on the balance sheet but if there’s a new crisis the Mandate requires central banks to absorb these risks on its balance sheet that no other institution or organization or company or financial institution will take during a crons and so uh sort of prepare for that and that’s actually what we did at the Dutch Central Bank which is most familiar to me we start thinking okay what could happen in in an extreme but plausible scenario now that we are in these asset purchase programs and so we use part of the profits because initially QE is profitable we use some of the profits to uh ACR additional provisions and capital now of course and added to that is so there’s no technical default but centl can default on their promises to provide price stability and so from a policy insolvency solvency insolvency perspective I believe that Central Bank capital is important now I think the IMF study was great so it opened a lot of discussion on on this I think CL got mentioned that or I think he referred to that back since then a lot A lot has happened so and it is very volatile so our projections are really interest ratees sensitive and so if there’s a small change in in future interest expectations interest rate expectations it has a huge impact on on Central Bank uh on Central Bank Capital now one of your questions has to do with um yeah so what is the impact or what are the the secondary effects of negative Capital so and I think there is indeed as Sarah also said a small impact on on the balance sheet but I think in the in the bigger scheme of things that are going on like normalization it has a relatively small effect especially in a world of uh abundant liquidity um then on the other question H so I think with in retrospective we live through a very extreme scenario uh H with a pandemic with a war and not every in every future scenario will be this extreme I would say and underlying um I would say central banks are profitable because of seniorage and so when all the projections that we have made we always see that in the long R longterm profitability the underlying profitability is strong enough uh to regain uh full capital but obviously this is very uncertain H so the world is uncertain there will be new crisis uh there will be new interventions by by central banks uh coming from the geopolitical uncertainties and the economic uncertainties ahead of us not only that Al also external developments like climate change loss of biodiversity um will necess necess new monetary responses and also cbdc I think that’s indeed what we discussed yesterday potentially can have a huge impact on the profitability of of central bank’s balance sheets um so that’s my maybe my first set of remarks thanks a lot what I get from here is no problem there may be some hiccups that we can all solve it so the future looks bright uh let me challenge this view a little bit uh and and the first thing is we don’t know what ASA is doing but if it has remains slow then the needs to use unconventional monetary policy will reemerge and this unconventional monetary policy in whatever form it takes place will bring us back again to situation what we have now not perhaps to the same extent because the dynamic there but we have it there the second part of it has to do with the question of scener uh yes sir with the reduction of cash this is lost yes scage in the future may increase but other forms of payment will stay on which don’t use uh necessarily cbdc so the Bas will be quite likely uh reduced so taking both together that suggest that the financial situation of central banks will how to say be reduced compared to what there were before and uh uh looking now what kind of major policies we could introduce in the past they will not be very helpful we to the central bank’s profit and loss accounts there and as result of it uh your your optimis seems to be a little bit for me overestimated what way happen in particular if and as one comes to the uh question of uh currently we have uh few well not many banks that have a negative Capital because we have an I think alluded to it at the beginning we had different Reserve levels there some banks had small some had large but also the larger ones will uh be diminished and then it’s not only becoming positive then to reincrease the reserve there was it won’t have the same level as in the past so but looks like you stick to it uh so uh uh some additional challenges to me or you say okay maybe but at the end of the day we will come up with some positive results so we don’t have to worry to it or do you see a need to think about alternative revenues you alluded to it alternative revenues but if and as Central Bank starts to use uner uh to uh uh use the money in order to invest internationally uh using the aner framework also is money creation so in this case you also have a pressure from this side as well and uh the possibility to make a lot of money uh given how to say how unfur has to be used to uh compensate that the European Central Bank are not small some additional ideas before we go to the money AR policy part which becomes then the uh reputational part any future any other ideas he wants to bring to the table or is this all my invention that uh that how to say I see many more challenges than opportunities to get back to a sustainable fin financing of central banks I can have a go sure yes please do so s it it I don’t think it’s maybe so much a question of whether the challenges outweigh the the the optimism um as it is a question of timing in some sense so uh this you know I I alluded to this question of the the long-term um financing model for central banks being something quite important that central banks do do need to take seriously some need to take it seriously very soon and and you know we we all know the central banks that are sort of if facing that challenge where cash is declining quite significantly Sweden New Zealand um and and so there I think it’s a question of what what is for for many you know many decades um central banks have had this structural um profitability because of seniorage income that can no longer be taken for granted so what’s an alternative way of of achieving structural profitability um you know you know Governors have hypoth iiz things around fee income for supervision responsibilities or uh or other things like that cbdc is is the obvious one as well um as an alternative to to private sector digital money but um so there’s lots of options on the table that need to be considered carefully I mean under no circumstances do I think central banks should get into the Mandate of profitability they’re not hedge funds they’re not they should never be um so so so I think it’s it’s not you know profitability for profitability sake it’s structural profitability um is is important to maintain in the very very long run but that has to be consistent with Central bank’s mandates okay uh yes please uh a maintain the same order um let me double down even more on the side of the purists right uh if if the IMF does not argue the economic truth which is Central Bank uh net income doesn’t really matter uh because of the uniqueness of central banks as issuers of the legal tender um then then who will so we we stay in that Economist world I think I also forgot my disclaimer in my first round but I’m not speaking for the fund or my current office I’m speaking as a co-author of this paper that said the policy lines we formulated did become our the imf’s policy lines for for the Euro area um we argued I mean we there’s three models out there for dealing with losses uh the reeks Bank model which is you know Central Bank goes to Parliament asks for recapitalization and gets it uh the bank of England model which is a standing Indemnity um and then the Federal Reserve model which is you have an accounting item you cumulate the losses you unwind it later when you make lo profits we thought the third was most appropriate for the Euro area you could again test me if one was in a situation of exploding losses would I say something different you know um and and if if losses and profits simply don’t matter does it follow that every Central Bank governor should have a corporate jet you know I mean so yes it is it is public money on some level um but the key thing really is you know to to us the confidence in the currency is what matters and that is a function of delivering on the primary mandate and that means to not be distracted by any uh you know to allow your policies to be colored by some desire to pursue some sort of profit objective and it becomes interesting because you know subsequent to our paper we had various little things happening in the Euro area I mean the 1% required reserves became unremunerated some national central banks decided to no longer remunerate government deposits um the announcement on reserve required Reserve remuneration introduced a sort of new word into the vocab the efficiency of monetary policy which I think was to be read as a nod to to uh this profit and loss Dimension um I have worried that this eurosystem might might choose not to do outright sales of security simply because that would realize valuation losses um so you know one doesn’t there are clearly risks so on some level this matters but it matters if we let it okay well this was clear I to say don’t worry uh State onment dat the rest is unimportant uh from a v perspective you agree with that that’ll say this is how to go ahead whatever High the carried forward accumulations are this possible to set the third part the third approach like the FAS bringing it Forward wouldn’t matter or maybe the need to change one of the provisions in EU law which states explicity how to say that uh at the end of the day one should become positive so one has to go back in principle unless is changed uh uh to the Mr Finance of Parliament ask for uh for topping up so your model May apply to the US but the current European rules I don’t think it’s consistent with if my if my if my assessment or my reading is correct shall I go yes please yeah so maybe to respond a little bit so first of all I don’t think it we should uh relax so I think the world is really uncertain one of the consequences of the huge losses that will take an awful lot of time before we come into a positive territory um and during this long period of time there will be new shocks to the economy for sure we don’t know which types of shock or Etc but there will be new uh shocks where central banks will have to intervene I think they also write papers on uncertainty and risk management so typically these are events which are not in our models and I think there other tools that we should use in Risk Management to anticipate that a little bit more and but still then the base scenario is in the very long term I think you really need to stress the models to not return to full capitalization but as I said there are huge uncertainties that will intervene with this like geopolitical uncertainties climate change uh and cbdc are the main drivers that will have an impact on Central Bank profitability going forward using anpa or investment is just a risk return trade off H so there’s no free lunch there it’s just a matter of are you willing to accept the the risks when these Investments lead to negative returns as opposed to senior R because that is sort of a monopoly for that central banks have and that they can use for their advantage I think a scenario in which many central banks will have a very long period of negative equity from an economic perspective seems rather extreme to me so this will mean that the term structure will be flat or inverse it will mean that we have deflation risk rather than inflation risk and that would actually mean that central banks are not able to deliver on the primary mandate and so um I don’t think that’s a very likely scenario economically speaking um and I think we all aware that that Central Bank should not print money to to resolve deficits and so in the end as I presented yesterday it’s what you negotiate or agree upon with the with the shareholders or typically the government either you do an immediately cash transfer in the central bank or you create this deferred assets which which means that the shareholder is willing to forgo for a very long time in for instance the case of the Netherlands uh future dividends uh and so in the Netherlands that is is acceptable so the Ministry of Finance at least the the previous one was willing to accept that um but I can imagine that in other countries or jurisdictions that might lead might lead to difficult discussions okay okay thanks then let us turn quickly to the second round because it was partly answered only because I would like to open up a a bit of discussion so uh currently as you said it doesn’t look like there will be a major pressure on profitability but if it were to be should it wor Central Bank about being capable of delivering on its mandate because number one uh it has to do that in the past uh the S banks have been a major not a provider of uh income for the Minister of Finance and for the time being it doesn’t look like in addition not only it doesn’t look like that we will give able to offer the money over the next number of years in addition there may be the need According to some rules to go back and ask for money for recapitalization in Europe out of it does this create a pressure on the independence of banks central banks so deliver on its mandate so that the governments are able to put pressure to exert pressure or should we also ignore this one because central banks can’t come under pressure uh let’s uh move for how to say no in the uh alternative round you start and then we move up here very quickly okay yes so um so um yeah indeed so at the point where it really comes to recapitalization you will have to open discussions with uh with the Ministry of Finance with Parliament with the government on how to do that and I think that’s a Terr that you ideally want to avoid the central banks because then if yeah politicians want to see something in return and I think okay yeah okay was quick and do the point you agree um actually I I I can offer the opposite view here so uh um If U if uh capital is negative and and no Dividends are in are coming in the foreseeable future it may actually increase in dependence because um there is small changes in current policy will have little effect on what the government can expect so it’s not obvious to me um what what the direction uh can be um but um yeah assuming that you can run constant deficits apped and that Dr back to in fact that Dr back to as long as the central bank has uh know uh Monopoly in the supply of money and and and at the same time um I mean we’ve seen in the data from our analysis that um the accounting rules and that allow as smoothing the budget constraint intert temporary and by building reserves and so on that that really help in reducing incentives uh but they they don’t exhaust everything there are other frictions Beyond the accounting rules so that they go a long way but there are in the data in practice there are other frictions whether they are political or behavioral that people perceive a negative number that something went wrong and therefore they see uh negatively interpreted negatively so yeah accounting rules help a lot uh but uh they’re not they’re not the end of the story okay um so I’ll quote my panelist and say it will matter for Central Bank Independence if we let it I think that was a great phrase uh so I I I’ll repeat it but um I think you know this this question it it’s whether needing that financial support is used as a lever to influence policy and I don’t think needing a recapitalization if that’s the framework a central bank has or a deferred asset if that’s the framework a central bank has would jeopardize Independence um unless it’s used to influence policy so how how can you avoid this I I’ll highlight a couple ways I think one is to avoid having to do it too frequently if if you’re a central bank um that needs recapitalization I think probably it’s better to do it every 50 years than every five years but you know I don’t think there’s an exact right number there uh but obviously the the the fiscal backing is there for a reason because even for central banks that hold Capital it’s probably impossible to have enough capital for for all states of the world in all time um and and so the other aspect I’d highlight is sort of the the to make it as automatic as possible so the rules of the game are clear um and you know whether it’s the Deferred asset approach or or or the recap approach um the more moving Parts there are the more risk there is that could could jeopardize Independence or lead to some kind of negotiation so having that clear framework more clear accounting rules for both distributions to the fiscal Authority and recapitalization I think are are the key so so recently I have been mischievously suggesting within the IMF that one could almost view Central Bank losses as a measure of Independence right I mean a central bank that was truly worried about losses presumably would raise rates less would be more distracted from its primary mandate so you know there’s something not unhealthy about these losses but um that was really just uh an aside I think the key issue again is balance sheet size you know and and and there are real moving Parts here um uh I think one need you know the ECB has been more adventurous than the fed and the bank of England in that it is countenancing the possibility of going back to a corridor system narrow Corridor could be widened later um but that’s an important statement it means they’re willing to uh face the kink in the supply curve you know um when but I mean I must be honest with myself if years if eight years of QE has put us in this situation it does mean we’re going to have a more animated discussion if we have to go back into QE again you know and many things in play one may want to within the Euros system visit the risk sharing and profit sharing on on bonds bought um and just one one other thing you know in the in in the realm of what is actually real for monetary policy let me give you one illustration the Central Bank does 5 trillion EUR in QE half of what it buys is from non-banks so it has not only created 5 trillion in reserves it’s also created 2 and a half trillion in Bank deposits you now have a very liquid banking system which does not need to compete for deposits it has enough so the Central Bank raises rates by hundreds of basis points the overnight deposit rate doesn’t budge transmission issue uh issue of uh you know the poor the poor man or woman who cannot afford to put his Savings in term deposits is hit hardest so you know there are lots of interesting questions around this and and I think uh there you know on the other side one could argue that there’s something stability enhancing about a system where as you raise rates the Central Bank gives a lot of interest income to the banks just as they’re going into a phase where their asset quality May deteriorate so one has to have a proper pros and cons discussion my own view is there are cases for a much smaller balance sheet where more stability on the p&l side is just a sort of desirable added Plus thanks a lot this last part was important because was a little bit missing that as we talk about the operational framework the question is under what circumstances is which framework more supporting of having a a positive balance sheet or not and I think you alluded to it and this is something what we have not discussed it in depths even at the central bank and I’m not aware that there’s a lot of uh not conceptual but also quantitative literature out there to make an assessment and what circumstances what regime is better which we have a bearing on our discussion we’ll have over the next two years I would assume so with this uh we have this part finished here would like now to open it up uh who’s the hands up there first David please well thank you David Marsh a fascinating discussion and I think a tribute to your bank Robert that you put on a discussion like this like the Ned luner Bank did also in April I do think that the secret to this is transparency and and talking about it uh that said I think the panel has been a bit complacent here and asok you were kind enough to say that maybe there are some underestimates in your paper of course you have made a rod for the back of the IMF not necessarily for Yourself by doing this superb piece of work because there will have to be a followup and as you say uh the estimates will no doubt be uh a bit more perturbing because interest rates have remained High I think your point is absolutely right this actually showed the independence of central banks up to a point because they’re going ahead with a fairly strict monetary policy even though it’s raising losses I would like to Simply point out what Isabelle Schnabel said in November 2021 where she said we can’t possibly put up interest rates uh at this stage before We’ve Ended asset purchases because we would be making losses so there’s unfortunately in the literature now a statement by Leading Central Banker saying that um losses do influence monetary policy and she now admits that she shouldn’t have said that or at least she shouldn’t have said it in a way that people might have read it but the the one question I’d like to put is the role of gold now when joim Nagle from the bundas bank said in end of February don’t worry uh we’re not going to have negative equity because we’ve got these gold reserves and we would not think even for a nanc selling any gold the gold price went up 10% in the weeks afterwards so is gold going to be if you like the secret Ally that will come to the rescue of at least some central banks the ones who’ have been sensible enough not to sell any thanks for the question I will suggest that we collect a few questions together with different View because they allow a dialogue not only a Q&A session assume thank you I’m glad that we have this discussion here um perhaps first a bit of a historical perspective in the Years let’s say 60s 70s 80s Central Bank losses occurred mainly if there was an appreciation of the currency of the central bank so when the Daymark appreciated against the dollar of course this had the effect that the losses that the dut bundes bank had huge losses with regard to the cadency reserves in dollar we have the same situation in the Czech Republic where they also in Slovakia where they have appreciation of the of their currency and that’s created losses this is the permanent problem of the Swiss Central Bank so this means that and it’s a strong indication that Central Bank losses in itself is a meaning meaningless concept it’s always the macroeconomic perspective that you have look to look at so therefore I think I should not get really get in the wrong in the wrong direction by comparing this with commercial Banks uh there is one element that would be a big mistake if there would be a legal obligation to recapitalize a Central Bank in the case of negative capital and this is for instance what we have unfortunately in Sweden it’s it is economic nonsense but this really is a danger for the independence of central bank because then the central bank has to start negotiations with them with the ministry of finances so so but we do not have this abs thir regulation anywhere else we don’t have it in Germany or we had didn’t have it in Germany we have it we don’t have it in Czech Republic and they working with a negative Capital was no problem so this is uh not not not not in a minute there is one point and Robert you mentioned this frequently I think this is very important of course Central Bank policy may have some distributional perspectives and this is with regard for just now the remuneration of Bank deposits with the central bank and where a large part of the losses That central banks have now is due to this renumeration and this means of course de facto a transfer or you could say a subsidy a public subsidy uh to the banking sector this is something where we should have a very iCal View and therefore I think it’s better sooner than later one should reduce this deposit rate and therefore also give room for a re reguvenation of the money markets but this is a specific aspect that is not relevant in general microeconomic thanks a lot a before I give uh the floor to somebody else just on this one yes the subsidi also have this arguments but more recently people made some comments their ways of looking at there their additional revenues but the banks have additional expenditures I couldn’t follow this argumentation but it has been raised if somebody could elaborate on it uh this would be uh important but the other part what you said you have this regulation about going back to the government in order to uh uh balance your uh negative equity my take is also in European regulation there’s the sentence that the encouragement not a clear request that yes you can for a little bit have a negative equity but at the end of the day you have to go back and ask for uh uh uh to to recap Li so just somebody knowing the law the legal basis better if they could answer to that Mr thank you very much for this extremely interesting and important discussion I have three short points the first one is I sense some inconsistency logical inconsistency among the panel’s prevailing views there are two views the first one is losses negative Capital do not matter for Independence and for policy vas you were the exception to this line of argument and the other view is central banks should have uh Provisions to recover from losses get Rec capitalized structurally make profits Sarah you you made this point and I think also DK so how how to square this why would the second statement be true if the first one were true my second point is very interesting overview by and and Ashok of the different Provisions regarding recapitalization among central banks worldwide now question is okay why is there the Swedish model uh the bank of England model and the FED model what does it depend on what is now the better uh is it history dependence is it totally arbitrary don’t don’t legislators not do what they are do not know what they are doing why are there these different traditions and regulations on the same issue and my third point is should structural profits structural profitability of central banks inform the choice of future operational Frameworks now I’m not saying it should be the prime driver but let’s say if there are two different operational Frameworks which are equally good for for achieving the Mandate of PRI of price stability should the central bank then choose that operational framework which is better for Central Bank profit structurally thank you thank you I think we have now enough of question not everybody should answer everything but let’s move in this round now okay good fantastic questions on logic I’ll come to those um I associate myself more with the gentleman at the back than with David’s argument that gold may be the Savior but he would have known that um I think an is is really raising interesting questions I I think the answer is as follows losses Central Bank losses Central Bank Capital doesn’t matter but this is not intuitive right it is it it really comes back to central banks being a unique animal and to fully grasp that the difference between a Central Bank and a Commercial Bank uh is not easy we and and so in a world of the ignorant um one does have the risk of pressure right you know a central bank that has spent decades telling its government to be fiscally responsible is now exposed because you know it can be accused of having bled and been irresponsible itself this is not true but it is something that may well be said what does one do one tries to take the politics out of it to the extent possible no you don’t want to go for recapitalization that makes headlines no you don’t want to go for money every year some people think that’s the best solution I would argue that makes headlines too um so sort of the truest solution to to to the reality of these losses not mattering is the one where you do nothing and by the way if uh you know I mentioned the dividend Interruption the there is a period of loss making there’s a period of recoupment and then you resume dividends at what level do you want your Prov your you know your recouping and replenishing your your provisions and and capital to what level I mean that’s a question that is fair enough to answer I would say um it’s good if you have enough that you know in most in plausible scenarios you don’t have to have an outright Interruption of dividends but that could mean fairly large buffers you know so I like that you pointed it out I’m just trying to uh answer it okay um so thank you to the gentleman in the back for for bringing up the the history there um we actually have a bis working paper coming out I think today um that that references this historical evidence indicating that a negative equity position um can be fully consistent with preserving ing the trust and money um I I think you know um you you mentioned the the the problem of the SNB as as you put it uh but of course this is a feature and not a bug in some sense um of the of the situation I think with regard to the Swedish example perhaps the problem isn’t so much the need to recapitalize as the need for negotiation to recapitalize I think that’s a really important difference and of course there’s some I mean you know they just got a new law very recently and the provisions there for recapitalization were not fully clear and of course have never been had never been used before until now um so that complicates things and in terms of the the the inconsistency I think I you know I fully agree with with your explanation I the it’s losses don’t matter because of the fiscal backing and the way that fiscal backing manifests itself can be in these different ways these different structures that you you you described so that’s not so much inconsistent as it is a a corollary perhaps um and that it’s important to preserve the trust and money that people know that the Central Bank does have that implicit or explicit fiscal backing when they need it at the end of the day um and yeah I you know the operational framework question I think I I think central banks should choose the operational framework that they think best um supports their policy objectives full stop um I I think it’s probably you know virtually impossible to say that they’re equally able to do that um you can argue which one is able to do it better and reasonable people can have that debate but I think probably to reach the conclusion that they’re both equally um able to achieve it is is unlikely um so so I think profitability should be a very very very distant second in that consideration I might even go a step further I’m not sure it’s fundamentally about fiscal backing it’s fundamentally about policy credibility you know yeah stop um just so going back to so slightly different but at the end the same conclusion in a sense so in theory um losses or negative Capital should not matter okay in practice and in the data it does okay so that that’s that’s what I can tell personally um um so then the question is what should you do okay and and and and I think that depends on whether you think it’s it’s even possible in that you Central Bank can be insulated by whatever frictions make it in practice to matter yeah so it depends on whether you think this is realistic that you are going to to be able to insulate the Central Bank from all these uh frictions um and um historically the data suggest that we haven’t been able to insulate central banks so I do think though that um in principle they shouldn’t matter uh and and and we shouldn’t let it and so maybe the best way to do that will be to recognize that they they will be in practice use or the frictions are present and then set up from the beginning a different sharing uh because eventually the central banks engaging in whatever operations and cure these losses for a a reason which effectively it’s it’s is is a fiscal um responsibility in a sense so uh maybe that’s the uh the way recognizing it from the beginning is is the best way and setting up the operations in such a way that that is is recognized last word ah okay thank you um a couple of points that will address many of the questions at least my views on on that so I think there’s first of all limit to Independence and so central banks are not fully independent and so Governors are sometimes appointed by politicians or by the government the central bank is accountable to Parliament at least the Dutch Central Bank goes every now and then to Parliament to explain for instance the risk of non non-conventional monetary policy we also disclose quite a lot on that in our annual report there is a annual letter going to the Parliament and the Ministry of Finance and also to public on the implications of monetary policy for Central Bank finances so and I think transparency is key so as as I said so I’m most familiar with the Dutch Central Bank we announced via a public letter in September uh two years ago that uh Rising interest rates will have a huge impact on our losses on our capital I think this was well taken also by journalists in the Netherlands and I think transparency can help a lot so risk management for me is also some times like communication and so we’re like a communication Department writing letters for the governor to the to uh to the Ministry of Finance uh so we have to explain very well the implications of monetary policy that we do it for the for the greater good that we have a specific rules in in rule in times of Crisis that involves more risk taking Etc I think we are glad that gold is there we use gold for the Very extreme scenario that we need to reintroduce a new currency so I think we would be reluctant to use it to absorb abor losses from other operations actually would also not be possible in the current setting without selling gold so the accounting rules prevent that so my view with Central Bank capital is important is sort of a insurance against political interference and uh my last point would be to look from from an exanti perspective on how much capital and Provisions you should have so it’s much easier to retain profits when you make profits than to do it after a shock when you have losses and because then it will take a very long time to become profitable again and to retain all the profits needed to recoup Capital thanks a lot I think it is a topic which we St started the discourse on and uh I claim it won’t be the last time we discussed the topic also because I see a number of still outstanding queries at least from my CER for example when you said well our central banks can have her how to say negative equity doesn’t matter there for me central banks and the Fisk are economically unit separated for reasons of governance but even if you separate them I to say uh the losses of ones are the benefits for the other way around so for me to still hold and this needs to be respective the other part I’m also surprised that for me the question is we moved from a system in which interest rate was the main instrument for monetary policy to something of unconventional interest rate so has to do with our star and the question is whether this change in which we’re in is totally innocuous for looking at our profits and loss accounts my take is it is not I to say it has weakened The Possession there which in my view has consequences then how to deal with it but these and other questions we won’t be solving even if we talk until midnight we don’t have midnight in front of us but midday and it’s time to have lunch and enjoy your lunch it is served behind you thanks a lot and I give you hand to this great P thank you yeah yeah good Robert hello yeah yeah this most an introduction to something which is part of my presentation uh to present uh FR Galo uh we only had two brilliant sessions this morning and I’m happy now to introduce a great colleague friend and fellow Central Banker fr Galo he’s now the second term as the French governor and also since 19 since 2022 he Prides also uh of uh the bank of international settlements so is the head of the supervisory uh uh uh supervisory Council which is currently now in charge of finding a successor of the current uh general manager Austin um he has been a main actor in shaping European monary policy with his knowledge and intelligence and if the pleasure here to have him here but they love to have him here in person but it’s greater that our electronic opportunities allows them to bring him here in full form here uh he’s not only a sharp C A sharp Sinker but an excellent Central banker and uh is very crucial in this changing world of central banks to guide us with inspiration but also with a lot of uh uh knowledge and experience there and uh so I’m happy for to have you here that you could join us today and as always for the reason of compliance as you have seen we see a lot of things there the one is Sir the mutual interest in the other languages foral speaks very well German and how to say so we when we speak we always alternate between French and German and in order to make us our language work the other thing is his uh his residency is in silent and there since I happened to be once professor in salent how do say the link to the this country University is also in common and I think we all both Europeans who believe very much into this important task there so with this small introduction for a big [Music] uh but but I will speak in English in English today and about this fascinating topic uh you chose for for this conference uh obviously central banks operate in a fast moving environment be from a technological economic or geopolitical standpoint and this indeed gives rise to Opportunities and challenges for the Central Bank of the future as an independent Central Banker I will obviously abstain I said from the start from any political comment today and we know we have a very moving environment in Europe and in France in this regard but I would like to focus on two aspects that are in fact correlated as we stand between economic soft Landing this will be my first point and growing public expectations and there is a link between the two of them let me start with an economic soft Landing which we didn’t achieve yet but which is inside we are gradually emerging from the inflationary crisis that has affected our economies and fellow citizens this since this ugly Russian invasion of Ukraine inflation peaked at 10.6% in the EUR area in October 22 it has since re seeded to 2.6% in last May uh then comes the question what is responsible for this fall in inflation obviously there is a reversal of the initial Supply shocks but monetary policy played its role it could be fashionable to say this is only by lack monetary policy did not have any contribution it’s wrong according to our own calculation monetary policy has contributed up to 2% inflation avoided this year along two channels the first one is a traditional one dampening credit and demand and the second one belongs more to Modern monetary policy keeping inflation expectations sufficiently well anchored monthly figures will be volatile this year due largely to Bas effects especially on Energy prices so we will of course continue to monitor actual inflation data in particular those on services but we will look still more closely at the inflation Outlook all the more since we have regained more confidence in our models and forecasting tools R and there were limited changes in our last quarterly forecast according precisely to our latest projections inflation should fall to an annual average of 2.5% this year 2.2% next year and 1.9% in 26 therefore as we are confident enough about reaching our 2% Target the time had come for a First Rate cut by 25 basis points and having a somewhat less restrictive monetary policy this is what we decided last week the deposit rate will stand at 3.75% as from tomorrow while the ne rate is estimated at between 2% and 2.5% in the Euro area therefore and this is important to stress at 3.75 we are still actively fighting inflation so 2% inflation Target is inside our efforts have been effective but but we are certainly not relaxing our resolve confidence requires vigilance as regards future decisions we are not promit to a particular rate path we will remain data dependent and I plead for what I call pragmatic gradualism both on the timing of future rate Cuts without haste no procrastination and on our terminal rate we have succeeded so far to achieve at once disinflation and a soft lending something quite unprecedented Thea has avoided the recession and remember it was feared 18 months ago with a positive growth of 0.6% last year and our projections expect growth to pick up to 0.9% this year 1.4% next year and 1.6 in 26 reaching our primary target of 2% inflation with a soft Landing rather than a hard one would be a lot better for the economy for our European citizens their income and their jobs and for the healthy conduct of fiscal policies so this is about the expected hoped for economic soft Landing but I come to my second item about growing public expectations toward Central Bank and there is a link between the two of them because the credibility of central banks and their Resolute course of action have a consequence they are fueling growing public Expectations first because central banks Inspire trust including from the general public to give you an interesting measure between last Autumn and this spring European citizens net trust in the ECB according to the euro barometer reach its highest level since 2009 and it increased in this last period in 18 out of 20 Euro area countries in addition to our first results in fighting inflation which I mentioned earlier let me recall the three components of what I would call The Golden Triangle of trust Independence accountability and mandate and say somewhere words about each of the three corners of this triangle most Central Banks gained independence from political power in the 80s and 90s including the eurosystem from its Inception Independence which I cherish does not mean however isolation the context of present uncertainities shocks and anxiety we cries more than ever that we provide to our fellow citizens and to the economy as much as possible explanations and clarifications some light in the dark if I may and this is about the second Corner accountability it is Central bank’s duty to explain for the monetary policy decisions in a comprehensible understandable manner including to the general public let us acknowledge it there is in our democracies growing distrust toward experts the right reaction to that is not for us to renounce our expertise but to renew our communication in an accessible and transparent way a monetary policy which is better understood is better transmitt and conversely an incomprehensible expert it happens an incomprehensible expert is no longer respected no matter how complex monetary policy can be the era of Cilan statements belongs or must belong to the past this is the first and most obvious growing public expectation but let me now turn to the third corner of the triangle the trust triangle our mandate price stability it’s a prerequisite for long-term growth and it’s rightly the primary objective of most central banks starting with our system by fulfilling price stability we contribute to our secondary objectives of economic growth social cision or the environment and let me here come to two other increasing public expectations climate change and coordination with fiscal policy how should we answer as the central banks of the future climate change first it has of course emerged as a major challenge over the last few years including for central banks given its impact on the economy and the financial cycle and on inflation and at we stress it it’s no longer only about long-term risks think of the DRS in Europe or of the floods in Brazil but also in Greece or Germany all had already significant economic costs as well as fiscal cost and financial cost I think for insurance the Euros system has strongly committed to taking action wherever possible within its mandate and has started doing so in particular through Market signaling the bank of France is also a spare head and has been recognized let me share this small Pride as the greenest central banks in the G20 by a panel of international NGS together with others we set up the so-called ngfs Network for Greening the financial system in Paris at the end of 2017 it now has more than 140 highly committed members including Liber Robert the Rish National Bank and we host in Paris its Global Secretariat the ngfs has developed I don’t know if you are aware of that seven long-term scenarios with a 2050 Horizon to shed light on what I will call the macroeconomics of climate change in all scenarios the economic loss in GDP is significant in 2050 according to other studies this economic loss could be double digit at the end of the century but all scenarios have also a common future the later the transition the more disorderly it is the higher the economic loss let me say one word about the next vintage of the ngfs scenarios to be published end of this year they will build on a new estimate of the impacts of physical risk which are greater unfortunately than previously F in line with recent advances in climate economics and climate science and hence they have as a consequence higher economic costs this estimate will meet the highest economic standards and will build on the expertise of the ngfs scientific Consortium be assured we central banks will incorporate more and more climate change in our economic analysis and forecast and in Greening our monetary operations Z said and I stress it that said it would be a monetary illusion to think that central banks can bear the bulk of the financing effort of the transition this is neither economically desirable nor legally possible green monetary financing would result in an inflationary Surge and it would contravene the European treaties which prohibit the monetary financing of deficits central banks and green Finance cannot replace sound public policies and corporate transition plans can I about transition plans we are making significant progress in Europe in parallel with regulation about public policies can I stress that one necessary step is in my view to put a price on carbon it would be in the long run the only signal capable of aligning climate imperatives with economic agents whatever form it takes and there are several possible forms carbon pricing will have eventually to be Global and not only National or european and socially Equitable I now come to the other growing public expectations in terms of coordination with fiscal policy the last few years have been marked by the Resurgence of Supply shocks as a consequence many countries sharply increased their fiscal deficit to Shield people and businesses from overly significant impacts can I sum it up this way this intervention was fully warranted for covid and it was you will appreciate the nuance partially warranted to counter the energy shock stemming in 22 from the Russian war but and this is the most important message what may first have looked like one of events now appears to be a structural Trend economic fragmentation along geopolitical lines and climate change May translate into more frequent Supply shocks either positive or negative and related public expenditures due to defense transition Investments insurer of First Resort could increase public expenditures by as much as 7% of GDP according to the IMF this change of paradig obviously raises challenges for fiscal policy monetary policy and they policy mix may I tell a small personal experience last September I had a friendly debate with Olivier Blan at the Paris School of Economics during a tribute to Daniel Cohen another brilliant French Economist who passed away too early Olivier asserted that fiscal policy is more powerful than monetary policy in this context also I see the point Behind These arguments I draw a somewhat different conclusion let me briefly explain why and conclude with that the first Nuance refers to an experimental historical limitation of fiscal rules however desirable they are discretionary fiscal tools may in theory be both efficient and well targeted but in real life beyond a potential oversizing such as the one observed in the US in 2021 in the end it is often difficult to reverse support measures even when the economic cycle improves fiscal policy stand tends often to be subject to political dominance and not only in the us or in France public opinion in many countries demand at once Rising expenditures and tax cuts under this political pressure many governments have failed to reduce deficits quickly enough when conditions improve and as all advanced economies have therefore experimented a staggered and continuous public debt rise the low rate environment that prevailed until 21 and the loan for an easing of the public interest burden was the exception not the rule and should never be taken for granted the second Nuance is more positive in Praise of monetary policy even at the effective lower Bond it remained very powerful thanks to unconventional tools whether accommodative or restrictive monetary policy has a clear quantitative Target the 2% inflation and it has several channels at its disposal we could see it in the unprecedented sequence of deflationary covid and then inflationary War shocks in the last four years whatever the circumstances monetary policy has found the right tools to achieve its medium-term objective of price stability not more monetary policy does not pretend should never pretend to be omnipotent but not less and this is not a small promise let us never forget that price stability is a necessary condition for growth and that price stability is the best way to fund Investments at moderate long-term interest rates on the long run hence we need both fiscal and monetary policies acting with clear objectives and full Independence but we also need the right policy mix it’s sometimes easier Rising t them as during covid and the accommodative policies on both sides at present the recovery cycle which is set to accelerate over 2526 together with a gradual easing of monetary policy create a favorable context for structural fiscal consolidation this window of opportunity must not be missed let me conclude and let me conclude with the famous French writer Michelle de mon who wrote his essays as early as the 16th century he was not a specialist of monetary policy but Robert as you speak perfect French I will quote in the original language of mon and then Translate de in Modern English it is doubtless a fine Harmony when doing and saying go together the central banks of tomorrow must say what they can do but also what they cannot do and they must also say how they are going to do this this is about accountability in particular with regard to our commitment to bring inflation down to 2% under these conditions I believe that central banks have a bright future because by combining saying and doing they contribute to the rarest but most necessary commodity of this Century at least of this decade trust thank you f Tran thank you for your attention and if there are one or two questions I would be delighted to answer them now feel mer thanks Dan for this great speech about a topic which in one the first part is easier about the soft Landing which we think the achiever the second part is a bit more tyy as it comes to the balance and the action of monetary and fiscal policy and as it look like at the moment we at the Central Bank seem to be on an easier case and course than the fiscal policy is but I’m sure there are people in the room who would like to ask the one or the other question I the one is nner and Mr Nik mer govern um I this morning I got a question could you because I I don’t see you could you introduce yourself please yeah sorry sorry Big N Austrian Central Bank working with Mr Holzman um I got a question this morning from a financial Market audience saying why should we believe central banks which are loss making what should I have answered uh this is about credibility and Trust uh and credibility in central banks should not depend must not depend I clearly think don’t depend from the financial results they should result from the compliance with the Mandate which is price stability let me let me stretch it on the financial results our financial results it’s true everywhere not only in the Euro area we influenced by the monetary cycle but we don’t run monetary policy to maximize our financial results or to limit today our financial losses we run our monetary policy to deliver on our price stability mandate and our 2% inflation Target and hence we must be financially solid enough solid including in our balance sheet to run any kind of monetary policy which is needed we perfectly know the explanations again for the monetary cycle and its consequences on our results there is no specificity of the European monetary policy regarding that but I clearly think uh if I look at most of other countries in Europe and elsewhere that this is well understood and this did not affect in any measure the trust of our fellow citizens uh in central banks I quoted and I think this is the best answer the measure by Euro barometer showing that the net Trust of Euro area citizens in the ECB and probably in each of it NCBS be National Bank bang of France or others is higher significantly than six months ago and at its highest point since in the last 15 years since 2009 so I really think there is our credibility relies on our mandate and our results for Price stability not our financial pnl I hope I answered your question thanks for SW we have now Mr Nik uh fans Naik European task for carbon pricing uh you mentioned that the price of carbon is really essential for achieving uh the green transition our problem in Europe is that we have reached the carbon price of around 10000 Euros a done but this has uh gone down to nearly 50 a done so it’s a very volatile price which makes it difficult for example for companies like toal which have an internal carbon price of around 100 in their investment models and also for a lot of other companies wouldn’t you agree that the European commission should like central banks when uh uh the exchange rates are too volatile and uh creating problems that they should intervene uh this is a very subtle and sophisticated question and you obviously know you know more about carbon pricing and its technicalities that I do personally So my answer will remain modest obviously we are not at the end of all the question about carbon price what form should it take should it be quotas like it is more or less at present in Europe should it be ATT tax as there other mechanisms how should it be compensated for perhaps some smmes and if I think to households for the more modest and there are very serious political problems there if you remember we had this move of so-call yellow jackets Leon in France six years ago uh and it’s more or less the case everywhere and the most tricky question is the question of international comparison uh because if you apply it only in one region Europe is not the only one but we know there are very significant exceptions then you have the question of fair trade should we have carbon border adjustment mechanism and so and so on you added one more question which is about the volatility and here I have my technical limitations to be honest my guess is that the more progress we make on the other question and the clearer long-term view we give about carbon price how it will evolve in the future uh is it stable for the decades to come will it be adopted by other jurisdictions the less volatility we will have but there could be a need for stabilization mechanism I’m not sure uh I’m ready to Guess that my friend Robert shares the same qution I’m not sure it’s the role of central banks it’s not in our mandate today perhaps some mechanism of stabilization could be used by other authorities uh independent authorities but I guess it’s only a part of the many questions we have my only message and here I concur I guess strongly with you is that as economists dedicated to the long run we cannot ignore the power of economic signals there are other ways of making progress against climate change there are regulation think of the green package of the EU there are subsidies think of Ira in the US and elsewhere but ignoring a a a price mechanism would in the long term means that we are economically not consistent I think it wouldn’t work by the way there are also interesting proposals by the IMF about a global carbon price with different levels according to the degree of Economic Development different flows so I don’t preclude this debate my only strong message is don’t put this item off the global agenda uh thanks for uh I fully agree with your assessment I think a carbon price particularly if there’s a price development uh projected well in advance will do most about climate change or because this is what investors neither I have a final question Robert if I may uh am I right to think again I’m not such a specialist but there is probably a link between the uncertainty about the future developments and I like the way you presented it to give predictability to the evolution of the carbon price and it present volatility there are too many uncer entities at present this would be my take my take is I to say a well pre-announced carbon price pass which may be adjusted over time what’s do most about it in order to make uh the green transition reality I have a further question here from David Marsh David thank you V thank you fris David Marsh here may I put a question to you about Capital markets Union I know this is something that you’ve given a lot of attention to over the years and you said earlier that central banks have to be aware of what they can do and what they cannot do could one of the things that you could do together with your colleagues on the European Council European Central Bank Council um make um a new stride forward in the onon cordal with the UK there are strong rumors that there might be a new government after an election coming up in the next few weeks could this not be an occasion where uh yourself and other colleagues could organize a kind of relance with the UK on technical matters which are within your sphere uh one can think of all kinds of technical ways of taking Capital markets forward without and having to rejoin the EU and I think if people like you and Rob were to put your energy and your muscle behind that it would have a lot of effect in the UK so David uh coming from you I’m not surprise that it’s a very smart question uh let me try to have an honest answer first we all in the governing Council the unanimity of the 20 Governors plus the six board members support an acceleration of capital markets Union within the Au and we took probably 20 bit unnoticed but uh let me stress it we took on March the 7th a very strong statement in favor of a new Step uh an acceleration of capital markets Union and we said many things I say it because we were perhaps a bit too modest in communicating around this statement but you will find many powerful IDs including a rebranding probably savings and investment Union if I follow my friend Enrico Leta and focusing on technological innovation including perhaps a unified Ledger green securitization Equity financing Etc four or five powerful levels uh instead of the 36 technical items we have in the present commission action plan by the way as you are aware one month later the N report was published in France with consistent IDs so there is a possible momentum also if I look at the European Summit uh to put it in a nutshell I remain in the EU at this stage of my answer to put it in a nutshell um it was seen as a very technical project which is a long list of items it should become a political priority because this is the way to use our hidden resource in Europe which is private savings we have a significant excess of private savings to fund our investment needs in green and digital transition and the amounts are very significant about the link with the UK I can only repeat the regret that the UK exit the EU but we we cannot repel brexit it has been a democratic choice so we will see with the new government uh if there is one in the UK with new governments throughout Europe let me you will understand why I am very cautious about any future political prospects we live in democracies we have electoral Cycles uh if there could be some common acceleration across the channel fair enough but let me stress that first and foremost we have the ways and means to accelerate within the Ura and and the EU uh so I I take these two flaw in the right order I think but you are right uh it’s part we are not the deciders for most items of savings and investment unit for some we can help but we are powerful voices to say this is common European interest and as we have limitation on Public Funding for obvious fiscal reasons let us use better our private resources thank you f we almost finished I would like to finish with a small question which is very much on my mind uh when we look into the interaction Monet fiscal policy what Rec currently have is that on the fiscal side we have for each cycle an evergoing increase in the level of that which we have in the Euro area there and uh this happens at the moment when we are coming from a period of low interest rates now the interest rates are going how to say uh uh we will be reduced them but they will still be much higher than in the past so how we able to break out how can we convince the fiscal sector that we or what can we do to support them in any kind of efforts to bring the fiscal house in order because I believe very much you know that we should deal with monetary policy they should do the fiscal policy but we see the the the the threat there and you mentioned it one way the Capital Market one way in order to relinquish a little bit of pressure for from them what are the other ways I mean you have been in in politics for many many years what would you say what would help us to help them uh this is a very important question Robert and a difficult one first I have not been directly in politics I have been in civil service so it’s it’s a bit different but let me perhaps share two Reflections I have no perfect answer to your excellent question but the first reflection is that monetary policy is run after its primary Compass which is price stability it’s not run after the idea of supporting or penalizing fiscal policies probably for the two of us it’s obvious but for some governments or politicians it’s less and for public opinion there is always a doubt remember when we decided very low rates and QE stronger QE during covid we did it to avoid the deflationary sh but many people were convinced that we did it to support Public Finance and this was not the aim and on the contrary when we had to raise rates and you are right for the future rates will be uh higher than in in the very ultra low environment we had after 2015 people criticize say you are penalizing fiscal policies you are not helping Etc but the answer is very clear we have one compass which is price stability and by the way if I look in the long run it helps to have probably more reasonable more predictable long-term rate so it helps but in a structural way Public Funding my second commment is what can we do we are not in the driver’s seat obviously and we shouldn’t be I think it’s very important what we can do is probably but differently in the different context what we can do is to try to explain to give the global economic picture and to say look in the cycle we are in probably the best policy mix could be if I look at present to use the gradual recovery we will have the gradual easing we will have to consolidate at a pace you will decide fiscal policy if I look at various central banks across the word there are different answers regarding this public expression I don’t know exactly Robert how it is for Austria in but if I look at the general rule for example traditionally the US fed doesn’t not say much about us us fiscal policy I guess what they think about it you guess the same we all guess the same but they public expression is limited let me express it this way because it’s a US tradition uh in Europe it’s a bit different and for a very important reason that we have one single monetary policy for the 20 countries but we have 20 fiscal policies so it’s quite logical to look at the various fiscal National fiscal policies and for each National central banks to say is it consistent or not with with the single monetary policy but to give the bank the France case and again there could be nuances when we say something when we write something and to give you an example I have a yearly letter to the president about our analysis of the macroeconomic situation I usually say some things about the macro fiscal mix what is the level of deficit what would be is it time for an expansion or a consolidation and at present it’s more for consolidation but I never enter into detail saying if there should be a fiscal tightening it should be X on reducing expenditures or why on increasing taxes this belongs to the political debate with government and Parliament so it’s a somewhat subtle border if I look at monetary policy we are in our Cor business and here we are independent if I look more broadly at economic and fiscal policy we are not in charge but we are allowed to speak and give our expertise and others will then decide this would be my imperfect answer to your excellent question no thanks a lot it was an excellent answer to my imperfect question uh because what you did is a man read with pleasure your letter to the president unfortunately we don’t have have a similar party in Austria so I’m I allowed to write to the chancellor to say something about fiscal policy I won’t do it but I it’s interesting Robert I must say that in France it’s by law 1945 so all almost 80 years now uh so it belongs to the Republican tradition okay now I I read it and I know it you also have in the UK so at beginning I was thinking whether I should do it but since I don’t have a legal opportunity I decided not to do it in any case thank you much for your excellent presentation and also answers to the question all the best to Paris and uh uh see you soon in I would assume in CRA uh uh for our next uh intellectual exchange all the best of B thank you [Applause] welcome to this policy panel number three on the theme monetary policy in the PO crisis New Normal let me to start with a catch up on something which I wanted to do yesterday but which I sort of you know got interrupted nam I wanted to thank the UN Governor Holzman bit nna Maria valderama and and their teams and also the event management team for your great coop for the great cooperation in setting up this conference this is really a smooth and excellent process thank you so [Applause] much in this session we are going to ask how have businesses consumers and wage Setters Behavior changed in response to the return of inflation the sharp rise in interest rates Rising uncertainty and multiple secular changes what are the implications for optimal monetary policy if we dare to talk about optimal monetary policy after yesterday’s lecture for for instance we have three senior Top Central Bank experts with us today to discuss these questions I start on my right with fella de Fior who is a research advisor at the monetary and Economic Department of the bis welcome fella to my left uh welcome to Frank smz longtime colleague advisor to the executive board at the European Central Bank and also swear fellow welcome Frank and uh further to my left mikal Horvat executive director monetary policy and market operations the National Bank of Slovakia uh we are starting out with three presentations of 12 to 15 minutes in this order then we’ll enter into the discussion among the panel with the audience we will start out from the from a Global Perspective with fella then get to the UR area ECB view with Frank and finally mikal uh will zoom in on potential implications from nonlinearities in price and wage setting for optimal monetary policy without further Ado over to you fiorella oops so thank you uh to the organizers for having me in this uh panel very happy to be here and uh uh Ernest gave a very interesting questions to uh uh address to this uh uh panel if I may um uh summarize them uh the first is how the inflation surge affect the wage and price setting behavior of Agents of private agents and the second is what are the implications for monetary policy from recent developments uh structural developments of the economy and uh of course let me remind you that all the views I’m going to express are mine and not those of the bis so I’ll address these two broad questions but from a somehow narrower perspective first I will highlight um one key element of the price and wage setting uh decision uh which is the formation of inflation expectations um and then I will move to the implications for monetary policy of some structural developments that we are seeing currently and these include the possible a larger role of supply side factors looking ahead uh in business cycle fluctuation and this may be related to uh uh energy transition but also to Rising political tensions in the global economy I will also look at the possibility that we will uh have economies characterized by steeper Philips curve uh possibly to do due to um increasingly tight labor markets and changes in the workers bar gaining power um and I will also uh consider the possibility of moving to higher levels of the natural rate of interest uh possibly because of uh age related fiscal pressure be there pension expenditures Health expenditures but also higher investment in green transition and defense uh and all I’m going to talk about is will be based on a joint work with colleagues at B um Ben Moon Dan re and Deano sandri um so let me start with the respon of inflation expectations duing the inflation Surge and let me first remind you of what happened here I have this uh uh CrossCountry perspective so I’m showing for six major advanced economies in red the Dynamics of a headline inflation in Blue Core inflation and in yellow uh the policy rates uh and the black vertical line is the point in time when has exceeded the 2% which broadly is the target for many of these central banks um so two uh messages arise from this picture the first is that uh uh the the this was an incredibly synchronized inflation surge um and the second is that the liftoff of policy rates occurred well beyond the point when inflation Rose above the 2% uh uh Target except for uh the notable um case of Japan uh and this uh also show uh visually the point made by athanasius yesterday of a period an initial period during the inflation surge of negative real rates um what happened to inflation expectations uh here I’m plotting the one year three year and five years ahead inflation expectations from the survey of professional forecasters um you can see that the one year ahead inflation expectations moved very closely with the realized inflation perhaps not surprisingly but the three and five years ahead expectations were much more stable um and um they they just Rose a bit more in uh uh in the Euro area which is shown here with the blue line and in Japan which is shown uh with the black line so the question that we are that I’m going to ask first is whether uh during the inflation surge we saw any sign of the anchoring of inflation expectations and we do something very simple we just use this regression where we have on the left side inflation expectations from a professional forecasters uh over three and five years ahead Horizon and on the right side we have a realized inflation we have quarterly data and we just run this uh very simple regression over the period since inflation exceeded the 2% threshold um and the estimate of this coefficient data is what is reported here with the red bars um and uh you you can see that the coefficient is uh um is positive and significant particularly ularly for uh the three three years ahead Horizon uh so there is some sign if you want of a movement towards the anchoring but the the the the coefficients are really uh the levels are really small so they they imply that an increase in a 100 basis point inflation uh in inflation uh led to three to five basis point increase in three years ahead inflation expectations so relatively uh tiny uh if you compare across countries uh uh you see that uh the positive and significant coefficient arises for uh us Euro area and Canada um which were also the countries where central banks had just undergone a revision of their monetary policy framework so and and among these the Euro area had a slightly larger coefficient so the next question I’m going to ask is whether um the the revision of this Frameworks which you could interpret as bringing uh a a a slightly more ACC accommodative monetary policy St on average in these three countries whether this this revision of the Frameworks influenced the anchoring of inflation expectation so we uh run a similar regression where now now introduce an interaction term so here this uh uh t m m mpf variable is a dami variable that takes the value of one uh from the time when each of the country embarked in the monetary policy framework revision and here we’re considering us Euro area Canada because these are where the main um countries that revise the framework and we run this regression over longer sample 2003 to 200 23 and we find that the the estimates of this gamma coefficient are non significant for us and Canada they are significant for the Euro area but they’re negative so our interpretation of these results is that the revision of the monetary policy framework did play a role in in the Euro area but actually it was a positive role it helped inflation expectations at the three years Horizon and five to re Anor towards the 2% inflation Target so overall the the the broad message from this part of the presentation is uh that that inflation expectations were very well anchored during this incredible uh fast inflation Surge and they were not effective by um or they were um they remain well anchored irrespective of the uh revision of the monetary polic framework just done by central banks so let me move to the second part of my presentation which is about the implication of post-pandemic structural developments for monetary policy uh here of course we need to resort to a modelbased scenarios uh to to be able to say something meaningful we use the DSG model that is often used by the new or fed for policy analysis for the us we estimated over the period uh 84 to 2019 so we stop prec covid then we take these estimates we use observables on um observables on the other variables and we then uh back up the shocks for the postco period and not surprisingly the shock that we back up are more um driven by Supply factors than preco uh and then we run stochastic simulations under postco shocks so more pre prevalence of Supply factors uh a possibly steeper Philips curve and a higher uh level of the um natural rate uh we consider two alternative mon policy rules the average inflation targeting that the FED adopted and the uh inflation targeting and the simple rules are just they just have a a persistent coefficient and the reaction coefficient to inflation Pi F and uh to the output Gap uh Pi F Gap and when we look at welfare we use this uh um simple loss function that penalizes deviation Square deviations of inflation and output and the policy rates from their natural level so we do uh three things the first is look at implication for monetary policy from a higher incidence of Supply shocks uh here what you see what I’m plotting is the welfare losses for different combinations of the reaction coefficient to inflation five pi and to the output Gap five Gap um under the it the inflation targeting framework uh if when the the shocks are uh distributed according to the postco um period so when there is more prevalence of Supply shocks and markup shocks and relative to the preco so there are two messages here the first is that under the postco type of Supply uh factors uh there is more prevalence of tradeoffs that monetary policy faces and therefore there are on average higher welfare losses so the the surface the red surface which is the postco one is above the blue surface which is the pre-co one so more losses for the same conduct of monetary policy when there are more markup shocks and and these also call for less aggressive for a less aggressive reaction to inflation the output gap which you see from looking at this uh uh uh Contour the black Contours so if you become more aggressive in the response to inflation to Output you manage to reduce the losses less if you have Supply shocks under average inflation targeting similar considerations arise but there is a more aggressive a need for a more aggressive reaction to output and why is that the intuition is that if you are um responding a lot to inflation like on this uh black Contour uh and you have a lot of Supply shocks you are uh possibly generating a lot of output volatility exactly because of these incidence of Supply uh shocks and so uh if you want to uh guard against this excessive output volatility you should increase your reaction to Output um what about a steeper Philips curve here we do the same thing so we take the post um post uh covid shocks more Supply shocks and we compare the flat the case of a flat Philips curve the green uh surface to one of a a steeper Philips curve the blue surface and uh the message here is that monetary policy become more powerful when the Philips curve is steeper because it has more ability to influence inflation and so on average under a steeper Philips curve the welfare losses are lower so more power for central banks at the same time this calls for a more aggressive response to uh inflation and the output Gap now under average inflation targeting I’m not showing it here the uh implications are very similar with the only difference that you need for less aggressive response to the output Gap and this is exactly because you have Supply shocks but your monetary policy is more influential on uh stabilizing inflation so you create a less output Gap volatility so you don’t need to react to Output so strongly the final thing I want to show you is the implications from uh a higher natural uh rate um for monetary policy here I’m comparing the losses under the average inflation targeting and inflation targeting so if you see a negative number in this table here uh it means that this is good news for the average inflation targeting because it’s lower in the losses more than the inflation targeting role and uh uh if you look now the first column gives the level of our star then the second whether Philips curve is flat or steep the third whether the shocks are more Supply or demand uh then you have the frequency of the elb and then the volatility in inflation output and the policy rates so for low levels of our star you see that the numbers here are all negative so that means that average inflation targeting uh achieves a better stabilization of both inflation output the interest rates but as you and and this happens irrespective of whether the Philips curve is flat or Steep and the shocks are pre or postco but as our star increases then you start reduces the frequency of the elb and also reducing the benefits of average inflation targeting compared to inflation targeting the numbers get closer to zero and when our star is sufficiently high it can turn positive at least for output volatility and that’s exactly because we’re an environment not only of high incidence of Supply shock but also varar being high so the benefits of AIT are low in terms of reduces the DLB frequency and higher in terms and the the costs are higher in terms of output volatility so that’s what I wanted to show uh main conclusions are that inflation expectations have been pretty well anchored uh despite the unprecedented surging inflation and irrespective of the monetary policy frameworker Visions um higher incident of Supply shocks increases the tradeoff for monetary policy call for a less aggressive response uh but under AIT the output response is key to guard against excessive output volatility uh under a steeper Phillips curve there is a more effective monetary policy so you can afford a less aggressive response and a higher rstar reduces but maintains but reduces the the benefits of AIT compared to it but uh if our star increases sufficiently um AIT can can induce higher volatility of the output Gap thank you [Applause] thank you fella for um flashing out how some key elements of uh a new poly crisis New Normal might affect monetary policy I think um you you you you emphasized three key elements and um I think Frank can build on this and I’m confident that Frank will also challenge some of the uh scenarios that you pointed out here Frank over to you I’m not sure about the last part but uh in any case let me first of all thank anes for the invitation it’s a great pleasure to to participate in this uh panel and and more generally in in this excellent conference um and of course let me first of all say the the usual disclaimer uh views expressed on my own and not necessarily those of the ECB uh or its governing Council so I I will actually focus on on I think a theme that uh already was mentioned in the introduction of the governor and then very much also was uh put forward in in the keynote presentation by uh atanasio uh being sort of the natural uh uh interest rate the natural real interest rate and sort of the uncertainty around it and and again fella also uh did some exercises around that I mean I will argue uh that uh we know that that uncertainty is is is very high and and we’ll show you just some indic ators uh but also that uh many of the structural changes uh secular changes uh that we’re talking about uh basically increase uh that type of of of uncertainty and particularly I’ll refer to some recent uh research uh with Jean gomuk and and G verand where we look at climate change scenarios and how different climate change scenarios different mitigation policies may also affect uh the natural rate as as was pointed out also by uh Governor v um and then of course the natural question is uh how do we design monetary policy so that it’s robust to this very uh big uh uncertainty and I’ll I’m very much a fan of of the work that atanasio has done on this and on sort of using the simple uh n growth targeting rules as as a benchmark but I will argue at the end that uh also the communication the current communication of the ECB which is in terms of three elements in the reaction function being on the one hand the inflation Outlook uh on the other hand looking at the Dynamics of underlying inflation and the third element being the strength looking analyzing the strength of monetary policy transmission is also a way of robustifying uh monetary policy in in the face of that uncertainty so that’s that’s kind of the the um the outline of my uh uh uh remarks um so I mean just the first part I mean we we know that uncertainty is very high I mean just to to illustrate that I I took two quotes and of course I could use many many many more uh by two uh famous uh economists one uh is Mori opelt I mean if you you read that quote this is from a paper at the end published at the end of the last year and it very much also I think reflects sort of the views uh within uh within the IMF uh where he says basically the main underlying factors that have pushed real interest rates down since 1980s and 1990s notably demographic shifts lower productivity growth corporate Market power and save asset demand relative to supply do not appear poised to strong enough to drive a big and durable Rising Global real interest rates and if I understood Governor Holzman uh yesterday very much he’s sort of pretty much also on on that line uh that contrast quite a bit with for example what Bill Dudley uh said only last week in in a Bloomberg uh editorial um so he’s the former president of New York fed where he basically said there is a strong case uh that our star has risen substantially uh because various factors are driving down desired saving and boosting uh desired investment again let me not read out uh sort of the details of of that that comment but basically he points to know some of the increased uh investment needs uh Capital expended that that needs to be done uh for uh renewable energy and also for um uh Green Technology techologies and and so that’s one of these secular changes that that we have been talking about and again I think Governor V made also a reference to that so very different views amongst uh very reasonable uh uh economists uh and so he says uh dley says put it all together and NSTAR could be as high as 2% again there’s also issues here I think of the concept is the concept the same maybe Mor opsal is talking more about the very long-term uh star maybe Bill Dudley is talking more about the short-term rstar but in the end it it’s it they both have in mind some type of stand or Benchmark for uh for for nominal interest rates once know the shocks to the economy uh wash out and and of course also uh consistent with price stability now what what does what are sort of the estimates of the the long-term neutral rate in the in the Euro area I mean my preferred very simple uh estimate is uh basically on the left hand side it’s just looking at fiveyear uh fiveyear uh ahead real uh interest rates uh this is the red line and so we show here from from basically the global financial crisis when it was the start was 2% it has come down uh all through the aftermath of that that crisis reached no uh bottom which was close close to minus 2% so we have here A variation of 400 basis points and then since the the tightening cycle uh this uh real interest rate which is no priced into financial markets has uh uh basically increased and is now sort of hovering around uh a little bit closer than a little bit above Z 0% so if you think about the nominal equivalent we’re talking about 2 to 2.5% which is ex ly also the numbers that Governor V mentioned so that that’s one estimate now of course you can do no there’s a lot of research uh with much more sophisticated methods uh at ECB we published a bulletin article uh sort of summarizing all that research that that’s here on the right hand side and I mean you see already just looking across methodologies there’s a big uh uh uncertainty a big range I mean at the end they all sort of seem to come around no 2% 2 and a half% so there’s actually quite a bit of convergence but I mean if you take into account that each of those estimates actually have has relatively large confidence sets I mean you have to put a relatively large confidence set around that uh that range um I mean one uh other sort of interesting uh fact is that of course since the interest rate tightening the volatility uh of bond yields has increased a lot and although it has come down it’s still quite a bit higher than uh no before the interest rate tightening uh cycle um so whether you look at at at Option prices or you look at realized variability and on the right hand side if you try to decompose know this fiveyear fiveyear forward interest rate into the real and the inflation component it’s mostly the real part this is the red component that drives this uh uncertainty and then I think this is a reflection of of also the fact that no central banks actually have quite uh big credible inflation targets so the the volatility of the the nominal components the inflation long-term inflation expectations is relatively limited what you see in the volatility of uh nominal rates is is mostly the real uh component so so also the uncertainty in markets I mean if you would compare the volatility of bond yields with the volatility of equity price is of course they also shut up uh after the interest rate and tightening and after the inflation uh uh pickup but they have basically come down to to pre pre- inflation uh uh levels so I mean just the a few uh elements to to to show how uncertain that is now basically the second thing is just I want to argue that many of those uh structural changes well thinking about democra trffic uh uh developments where we think about uh technology digitalization Ai and productivity uh but also when we think about climate change uh can have very uncertain effects of uh on the natural rear rate and again here I’m I’m I’m just showing and of course I can’t really uh explain very much the the details uh behind what we’re doing here and the model that we’re estimating but basically in this paper what we’re trying to do is we we estimate the very simple basically four equation Newan climate model we call it where we kind of extend the Newan framework uh with basic the basic elements from the uh integrated assessment literature which sort of models the effect of of uh global warming its negative impact on on on output and and on potential uh and also the implications for uh for mitigation policies and so we estimate that that simple model on world data because global warming is a world phenomenon so this is not Euro area specific it’s sort of a world uh uh model and then use it to uh to simulate different scenarios um again for uh the the estimation and the calibration of of the climate related part we we we basically uh use uh no the dicey literature the nor house uh literature uh what we show here is is basically the difference between um uh two extreme scenarios the red line is is basically what we call the less a fair scenario so so no uh policies that uh attack that attack or address uh global warming and carbon emissions so the tax rate no the favorite instrument of uh of economists doesn’t uh rise as a result in panel B you see that carbon emissions start uh increasing uh in our model that will have a negative impact on on output because of all the physical uh risks and and effects it has on potential output and of course as the the the uh the globe warms these effects will start increasing actually the inflationary effects will be relatively uh uh limited um uh and and here in panel F you see the impacts of on the natural uh real rate in this model and so you confront that with the green uh scenario which is basically our calibration of the Paris agreement which uh where the goal is to bring down the carbon emissions no Net Zero uh at 2050 so that’s panel B uh the way you do that in the model is by increasing taxes and as as was discussed uh in the previous session so the best thing is to have like an expected increasing path uh of the the cost of uh of of carbon um that actually leads because you increase taxes in the short run it has a sort of a negative in in output but of course the gains are really in the in the longer run you get a much higher output and you also see that there is this sort of Hump uh around to 250 which is related to the fact when you increase taxes no the reason why you increase taxes is to provide incentives for firms to Green their Technologies so they have to invest they have to make uh uh abatement efforts and of course that uses resources no and and so it it is a source of exogenous demand um but of course given that Supply is is limited what in this scenar you need is is also a rise in the natural rate because there will have to be less consumption to make F there a crowding crowding out so to say and so the whole point here is just to to to show that no we we just look here at two extreme scenarios that they have very different impacts on the natural uh rate and so what we do in the paper and and so this second scenario also have larger inflationary uh impacts now of course that depends on monetary policy and the monetary policy rule that we use in these simulation is one that has a constant natural re rate it does take into account an output Gap where sort of the negative potential output effects of global warming is taken into account but it doesn’t adjust the the ninal rate and you see you get more uh inflation the inflation effect can be very different of uh the Paris agreement scenario and here we just uh show two uh or three alternative uh rules I mean they’re all of the type of uh tailor rules with interest rate smoothing the the green one is what we call in our model the Baseline rule that’s the one that we used in the previous uh graph where the real rate is constant and you have no an appropriate output Gap um and so you do get a bit of inflation uh uh and normal interest rates will have to be be higher in that in that case if the central bank would follow a rule that just targets output and doesn’t take into account the fact that in this scenario there is a negative Supply effect of some of the warming on on Supply then you get no the possibility of a a big uh rise in in inflation or or much slower return of inflation which is is panel B uh to uh to the to the Target so you get volatile uh uh inflation if and that’s the the long run uh sorry the natural rate adjusted rule so if you think okay the Central Bank actually knows also the implications for the natural rate that I showed and it sort of takes that into account then I mean there’s not really an inflation uh problem coming from the fact that no global warming on the one hand lower Supply on the other hand increases demand and in principle should lead to inflationary par green inflation climate inflation uh what have you um if the Central Bank realizes that again then the inflation actually is under control and uh and so now of course that’s a big uh question does will the Central Bank know uh what potential output is what the natural rate is and I’m very much with aenasius on this that basically we don’t know and that was the whole point of also showing the the first set of of of of graphs and so then the question is and and this is where where I will end where does that leave us in terms of how dependent or how can we robustify monetary policy uh with respect to our estimates of of what the the the true rstar is what the true interest rate real interest rate is that balances Supply but savings and investment I mean why is that important because if you look at our Baseline projections underlying it uh and Baseline projections are the ones that justify also the most recent uh cut in interest rate underlying it is the Assumption though that the policy rate will uh uh sort of over the next two years come down to a nominal neutral rate of around two uh 25 2.5% that’s what analysts assume that’s what priced in the in the markets and of course that kind of assumption also uh is behind know the the assertion the the statement also by by by the governor vwa that our policy is indeed restrictive at in the current context and so on the right right hand side uh we calculate based on our survey of of monetary analysts what we call the real rate Gap so that’s basically what out the analysts think the real rate Evolution will be uh for give take taking into account their own expectations of output the noral interest rate and the uh the the the uh the the neutral uh ninal rate and so you see it’s it’s it’s it’s tight uh and and the question is what if and this is I guess the the question that also fiorella uh raised uh what if we’re wrong in or and again it’s not necessarily Central Banker it’s it’s the analysts no we kind of take uh that view that is in the markets as as a as a starting point for our projections um and again I would argue and I’ll stop here that uh the sort of the the three elements of of the reaction function that we uh DCB has has communicated so so one way is to do basically what what atanasio s suggested sort of use as a benchmark a nominal growth targeting rule which does not depend on an estimate of our star uh but just says no you raise interest rates if uh your expected nominal growth is higher than your estimate of the nominal growth Target inflation plus potential growth and you lower it uh when when when the opposite the issue with that so and I think that’s a very good starting point I’m a big fan of it the issue with that is that of course the that expectation is partly a function of what you assume the the nominal neutral rate will be so how do you make sure that also that rule is is is um uh robust to to our star uh uh uh uncertainty and and again so I would argue that uh the first element is basically the Baseline so so that’s the that’s kind of the underlying assumption that nominal interest rates and real rates will will go back uh to to to neutral rates the second element is really looking rather than the forward looking which is in the Baseline putting a bit more weight on the backward looking compon which is underlying inflation it’s wage inflation it’s domestic inflation so which are slower moving but the the problem is that these are backward looking variables and the third element is basically the strength of the transmission mechanism so if this view is right then actually we should see that monetary policy is uh tight no that it affects uh Landing that it affects uh interest rate sensitive uh components and that’s exactly what what we’re seeing no growth has been restrained and so by by by sort of confronting the Baseline view with some of those other elements I think uh we can also robustify uh monetary policy let me stop here thank you very much [Applause] Frank also for zooming in uh a bit further and deeper on the challenges for monetary policy from climate change topic that uh Governor V guayu started out and uh you really brought some important additional insights so we now um go to mikal uh for some considerations on nonlinearities please over to you um thank you and thank you for the thank you for the invitation good afternoon everybody it’s great pleasure to be here um yes uh I’m going to talk about nonlinearities and um when we talk about nonlinearities and monetary policy uh what immediately comes into most people’s mind is the obviously the zero lower bound which uh has been uh studied extensively um many of us have been uh producing or looking at pictures such as the one on the uh in the slide for for a decade U it’s there for purely for decoration so don’t read too much into it um it’s uh what we were doing for the for a long for a long time but in the meantime we’ve also had a period of um High inflation uh which um brought kind of new phenomena new con considerations the the main question in the for the panel was if this period has led to um some kind of different type of behavior by uh consumers and firms so um perhaps there are other nonlinearities that we should be uh paying attention uh to going going forward um the the title of the session is about poly crisis normal the way I’m uh trying to read that is uh that um we’re facing a a future where either um sort of bigger shocks can happen in quick succession or even coincide but it doesn’t have to be a supply shocks I mean it could be um could be innocent demand shocks in a in a in an environment where uh Global Supply chains operate um with with low buffers and so fairly small shocks could generate bursts of inflation and so the question is if there are any uh nonlinearities on the upside that we should be uh we should be thinking about and there’s a uh sort of emerging literature U spurred by the sudden availability of data from a from a high inflation environment about what what actually people did during this uh during this period and um so the main Narrative of my talk today is basically about uh uh where’s threshold rates uh so there’s some evidence that people started paying more attention to inflation in this uh High inflationary uh period um so beyond a certain rate people suddenly start paying u a lot more attention to inflation and also firms um start to reset prices at a at a faster Pace at at a higher frequency so the available evidence uh in this area currently points to uh threshold rates which are are pretty low certainly lower than uh I would have expected um so uh um that kind of uh made me consider this uh topic a bit more uh seriously put a bit bit more weight on these arguments that I otherwise um otherwise would have uh key question obviously is whether these threshold effects have any uh uh important Economic Consequences um and I guess more research needs to be done on that but there are there are some there is some evidence uh out there that um they actually do uh exist and affect uh people’s uh people’s behavior in an interesting way and finally um uh the question for the panel I guess is also whether we should take these into account in the design of um monetary monetary policy there I have to say I’m uh I’m I’m wholly unprepared uh I’m prepared to speculate but I’m uh this is a part where I think a lot more thinking needs to be uh a lot more thinking needs to be done before we draw any conclusions but um so let’s see how this goes today so about these uh threshold effects um so there’s a a paper uh coming out in restart U which um basically looks at how um people’s attention to inflation as measured by Google searches or Twitter data um varies over over time um relative to the uh headline inflation rate so the left hand panel replicates this uh this work on on Slovakian data and indeed you can see that people as as inflation started to rise uh people started searching for inflation a lot more than uh than before and when you uh do some U econometrics on that uh with threshold um effects you see that there are kind of two regimes uh there and the uh the threshold inflation where this um inflation rate where this uh search intensifies is actually um somewhere close to 4% for for Slovakia um the figure in this previous slide is from the same paper showing the uh situation uh for the uh for the United um States so basically for a large number of countries you have a um you have a situation where for when inflation is low people don’t care they don’t they don’t pay attention to it at all and then suddenly when it starts to rise they uh start paying uh more and more attention to it um it’s not only uh Google searches and Twitter data and these fancy new things but we’ve also looked at uh data from the um um joint harmonized uh commission survey among consumers uh later on I’ll I’ll site this database for businesses as well we have access in Slovakia to uh individual answers so we do quite do work quite a lot with this um with this database but even if you just look at the uh look at the aggregate um there’s a question in this survey about uh inflation um uh perception so how how consumer prices have evolved over the last 12 months and if we treat uh uh the don’t know people as inattentive and plot one minus don’t know as the attentive people uh the share is uh very nicely um very nicely correlated um so as inflation starts to rise people suddenly have views on uh what the inflation uh what the inflation has has been again this is um data for Slovakia but basically you can plot that for um any any EU uh country and um when we look at so the question is where where is this uh uh threshold rate for Slovakia it was just under 4% for other countries is different the histogram for a large number of countries is the following again taken from that coreno paper so and and then the median actually is uh is pretty low um it happens basically fairly close to uh um the um the the targets that most most economies uh have obviously in some countries it’s uh it’s larger and you obviously would be asking uh whether what this is being linked to and it is linked to the experience with uh uh inflation over longer period of time um which is which is this picture so um it it comes under the headline whether this matters at all um um and and the main message here would be that uh uh when the experience with inflation is uh is higher it uh people are kind it becomes a part of life and so people uh um Talk notice inflation a lot more than when inflation is uh low for some countries with high history of high inflation rates there is actually no uh threshold um no threshold being uh being measured and basically the paper classifies countries into different groups where at low inflation rate basically people are completely inattentive to inflation there are some and Slovakia is one of those where even at fairly low inflation rates uh people notice um inflation um more and more and then there is the uh then there is the uh threshold now again so the question of whether whether this actually matters in in a bigger scheme of uh uh things I I guess the jury is still out there uh there’s a nice paper um uh by fouy uh very recent one uh who first replicates this uh estimation of uh inflation thresholds and then builds that mechanism into uh standard of new Kian uh framework and and looks at um what happens in response to a supply shock in uh in in such a framework and so if if there’s a supply shock that is big enough uh to push inflation above uh this threshold um what happens uh there is that agents start to update their expectations more strongly in response to uh the the supply shock um basically what’s happening there is uh that the um aggregate supply so the Philip curve becomes steeper and at the same time the aggregate demand um curve becomes uh steeper so a given shock is uh generating more inflation and is generating more inflation expectation so kind of this this this is an important amplification mechanism in that in that framework eventually uh as the shock dies away things calm down obviously but uh there’s a there is a nice mechanism towards the uh towards the end so as as inflation Falls because the shock has died out um people stop paying attention to inflation uh the thing is that their expectations May well uh remain stuck so they they don’t notice actually that uh inflation has come down as significantly as as it has so their expectations uh um get stuck at a higher level and this is so he proposes that as a um as a way of thinking about the last mile which uh we hear quite often um uh these days so interesting story uh question is is whether there is evidence um for each individual element in in in the story or to what extent this is we have we have corroborating evidence from Elsewhere on on uh these points um uh we are doing some work with the uh micro database as I I mentioned uh which among other things it asks people uh whether it’s the right time to purchase durable good uh at uh at any point in uh at any point in time and uh so our estimations my colleague shows that uh basically the answer to this question is linked to uh inflation expectations in a in a positive way across the whole sample so you can take different uh from these surve you can take different measures of inflation expectations qualitative or quantitative different kind but there’s a there’s a positive relationship between um people thinking that it’s it’s good to buy uh cars and washing machines and other durable goods uh and and inflation uh expectations it is there in normal times as well but this relationship is a lot stronger as uh inflation uh as inflation surges so that kind of um points to the fact that there is amplification uh going on um in times of uh significantly rising Rising inflation um now key element towards the end of that story is that uh inflation expectations get stuck uh as people don’t notice that inflation has uh has has come down U now for this the uh the evidence from uh these um European survey data is that actually expectations have started falling quite rapidly uh recently much faster than perceptions previously they were closely linked now they are um now they are falling um they’ve detached from perceptions which is an interesting question why they have I don’t want to go into that but um actually expectations luckily have been uh falling um at fairly fairly steeply recently the figure is for Slovakia but it works for um the whole uh pretty much all European countries and the uh and the Euro area as uh as well so I guess more needs to be more work needs to be done about the what happens in the as as inflation starts to uh um starts to fall there’s also uh um the fact that there are there can be other nonlinearities and it’s it would be misleading just to focus on um single one of them um and um another thing to consider uh when we think about the high inflation period is that higher inflation also brings uh more uncertainty about uh about inflation so here the picture plots uh a measure of uh uncertainty against the inflation rate again this surveys based thing uh the uh the the measure is uh the fraction of people who give round answers to uh um questions about uh inflation and and the share of people who say don’t know round answers are a a measure of uncertainty going back to a paper by binder and the in the GM jme uh basically people when people don’t people are uncertain they just tell you a round number so they would not say so it’s uncertain they would not say it’s 3% or 4% or 3.4% they would just say I don’t know 5 10 15 50 um so uh that that’s uh that’s what is comes from literature on psychology and uh so if we plot this share we can see again a link between uh uh between uh uncertainty and and the level of inflation and um it’s a well-known well I mean um in economics we tend to think about uncertainty as something that actually reduces consumption uh rather than um uh increases so we have one factor uh expected inflation perceived inflation um basically telling people to buy more and then at the same time uncertainty is going up and goes against again we have um evidence in in favor of that uh as well so we um we estimated um what what attentive but uncertain people uh do so attentive but uncertain people are people who um um provide a perceived inflation figure so they uh can put a number of what inflation was uh um in the in the recent past so they are paying attention to uh what’s happening to prices but when asked about um when when I asked about the the future they they they have they say I don’t know so and in this this category uh there is a uh a negative significant negative impact on consumption uh coming from uh coming from U increased in inflation kind of offsetting the uh impact of uh Rising um inflation expectations so um on many elements of that narrative the uh the jury is out there I think it’s an interesting area of research also uh working with these micro databases it’s uh it’s it’s quite fascinating stuff understanding uh um the uh consumer behavior um better um a quick few slides on on the firm side it has been well known for from the literature on emerging markets that uh uh the frequency of uh um price changes by firms is linked to the level of inflation that’s from a QJ by Alvarez uh which shows that um as uh the um annual inflation rate Rises the uh firm’s reset prices uh more often there’s this flat bit uh at the bottom which is where we uh think we operate most of the time uh but the uh recent High inflation period uh brought inflation rates where we might have just caught the uh the the the region of liftoff in this uh in this um in this picture so it is kind of legitimate to ask if there are any threshold effects that we we can measure on our recent experience as well when it comes to uh firms uh price setting and it seems that um there is so we can get a similar picture on the uh on the on the firm side again from using data from this uh commission survey firms are asked whether they expect prices to increase decrease or stay the way stay the same over the next three months and so we can plot the share of uh people who are expect either price changes or price increases and at least the in in retail sector the the link with inflation is uh is is pretty pretty strong and then you can also do the threshold estimation and again uh you get um evidence of that at pretty uh low rates of uh uh low rates of inflation so I guess my point is that these are interesting things to uh consider um going ahead if we are indeed entering a a world where um we can get bursts of inflation that can potentially last uh unpleasantly long um for from the perspective of consumers and and and and firms and the policy implications I have a question mark there so I this is more like a question for discussion whether um um this sort of narrative is telling us that seeing through uh Supply shocks as as the literature would normally recommend is it is it a risky strategy is it um something we should uh we should forget about and step in Fairly on fairly on when we see such Dynamics and unraveling uh it’s also a question about uh um the conduct of more system so there are questions about the conduct of more systematic uh monetary policy um um there’s a there’s a literature on uh uncertainty about the Persistence of inflation nice recent IMF working paper which uh basically argues in favor of uh more forceful action to uh basically um stabilize inflation more if there is uncertainty as as as robust policy in a in a world where there’s uncertainty about um about uh persistance to rule out some pretty pretty unpleasant um uh Dynamics um and then yeah so uh normally we tend to think in terms of inflation targeting but as fella talked about there are other other options out there um and and normally they um in the standard World they do quite well at least in in in models and in theories they do well in stabilizing inflation expectations um um yeah the question is how they uh how whether they would be a part of a problem or part of the solution in a world where um nonlinearities like this exist so thank you very much for your attention and for the questions thank you very much mik thanks for elaborating on threshold effects in the area of inflation expectations and price setting now I Take Along two figures which I wouldn’t have you know found otherwise so for Consumer uh expectations a threshold effect of 4% was shown by you something like that uh for Price setting of firms 2.36 or something like that very precise figure um so this is relevant in so far as it has implications for the reaction to supply shocks as you pointed out in your last slide and I found some interesting tension with your conclusions um uh you know in reply to supply shock so maybe we use a couple of minutes for a first round of uh reactions to each other’s presentations Fel yes thank you uh yeah very very interesting presentations both um I had a couple of reaction one to Frank and one to M uh on Frank I I find that uh you know the the point you raised about uncertainty and uh uh about our star is very is key so we just don’t know where it is and how does that affect uh uh the conduct of monetary policy and uh indeed we we did run an exercise of that sort assuming that the Central Bank just doesn’t realize maybe there is an increase in our star but the central bank doesn’t know and so uh how do alternative monetary policy rules um compare there and and so what we uh uh got out of this exercise is that you know you because the the Central Bank does not realize and it doesn’t then adjust the policy stands as sufficiently there is this period of overheating of the economy but under that circumstances the the overeating is less pronounced under an average inflation targeting type of rule simply because that rule uh promised to tighten in the future to set the overshooting um and so seems to perform better than the inflation targeting role that said the question is what is the type of uh role that indeed even an average inflation Target in Central Bank like the FED is actually uh adopting is it if it’s asymmetric and then the commitment is not there in the presence of i period of inflation then probably this mechanism would be uh mitigated in reaction to the last presentation and you had a question at the end how would alternative roles perform under this nonlinear Philips curve it’s a very interesting question we haven’t done that but my intuition would be that again the the the type of uh makeup strategy would work well why because when you are in the flat part of the Philips curve H this is where demand shocks can drive you to the um zero lower bound and that’s where the the mark uh makeup strategies um achieved most benefits in stabilizing inflation expectation minimizing the currency the occurrence of the zero lower bound um and when you move to the Steep part of the Philips curve that’s where Supply shocks might uh uh be more prominent but that’s also and that’s where uh the average inflation targeting would penalize because would create a lot of output volatility but it’s also the point where the Philips curve is steep and monetary policy is more effective and so that would uh again minimize the cost in terms of out volatility so my intuition would be that probably it would be even uh the results would stand there thank you very much Frank what’s your take on these points and what do you want to comment on the other presentations um also maybe two two comments I mean first on on um your empirical results on uh the the anchoring of inflation expectations uh I mean your findings that no initially in the Euro area case it was about re-anchoring I think is very much uh uh consistent with some of of of our own uh estimates in that because obviously at the low interest rate environment also long-term inflation expectations were somewhat below uh 2% um and so from that perspective no that’s a good uh thing and of course the risk then comes when when it rises above 2% but I think it does have also implications for for example the some of the evidence that atanasio showed if this is the policy and and the the intended impact from the the strategy uh review then actually you would expect that re rates have to fall no because you push up inflation expectations noral interest rates stay and so the real rates and that actually know becomes a self fulfilling uh equilibrium and so the fact that real rates fell uh towards the end of the the low uh interest rate period I don’t think is evidence of being behind the curve I mean it’s it’s could be interpreted as evidence that actually monetary policy and the forward guidance was uh was effective the question then is of course at what point do you switch no and and and again I’m not saying that uh central banks uh uh were were maybe not uh too late but one implication of of that of strategy is that then you have to raise once you start you have to raise much faster and I think that is one of the risks and that’s what central banks did but it’s also one of the risks particularly if you think about for example some of the financial stability uh implications and so I think that that kind of tradeoff uh central banks have to have to take into count um and the other uh I think it was it’s it’s it’s interesting to combine uh both of your uh presentations because if agents at low inflation say between 1 and 3% are not consumers particularly households are not very attentive then I’m not sure average inflation targeting which very much depends on managing inflation expectations quite precisely know you want to actually steer inflation expiration to be higher at at at the lower bount in order to get the may actually not be feasible so so my question to you is and we did some of that work also in the your system when we had our strategy review uh how do these results of sort of the benefits of average inflation targeting uh hold up when you have some of those in attention uh features in your model because I understand the model you’re using has a rational expectations uh uh model um so let me stop here okay I thank you than you Mikel first round of reactions and I think I need then to again give Fel the opportunity to reply yes uh so I guess the the part where of P’s presentation which uh got me thinking was exactly about this relative ranking of alternative monetary policy regimes uh my immediate reaction um when I um first started looking into uh uh these questions with the help of my colleagues was exactly that um uh perhaps this could be an argument for a some kind of a makeup strategy given that it at least in models is very effective in um in anchoring expectations and and and perhaps we would uh through that mechanism we would uh avoid actually or mitigate or sort of dampen some of these uh um inflation implications of of shocks that it should actually bring um um these topics um to the FL uh uh then I was uh my my immediate reaction was very quickly count down references to richer on uh on uncertainty shocks and and and the the some of the conclusions that is coming out from there about um being a conservative Central Bank responding strongly to inflation being more important so I guess um um you know this is one of those things that probably needs to be properly investigated in in models consider all the tradeoffs of or or the channels through which these alternative regimes operate before we can um be more confident um pointing in uh any of the directions uh from from Frank slight um obviously uh the uh um the link to uh my talk would be the timing of the of the tax increases so uh the you know interest increases tax so it was the shock was about shock was about carbon tax so when uh it’s it’s clear they have to go up in the in the long term but um you know they can go up smoothly or they can go they can have a steeper profile first and and obviously it has different implications then for uh for climate change and for what happens in the uh in the real economy but it has implications for inflation as well so if uh if there’s a steep price at the start which leads to uh increases in uh inflation if for example the rstar is mismeasured um then um then we could get into some of the troubles that um that that potentially arise in a high inflation world thank you fella you do you want to react immediately very quickly so first I completely agree uh research myself documenting the insensitiveness of uh particularly house households to monetary policy to to announcements um probably to the details of the strategy that applies to makeup strategies probably applies also to moving from an asymmetric to a symmetric Target it applies to the to the effectiveness of the transmission of monetary policy so it has very broad implications even beyond the makeup strategies and uh I know that at the FED when they did the review they did a look into using the FB us different assumptions about uh agents uh expectations and the rationality they found it was robust to some extent but yeah excellent so we have time for maybe one or two questions B please thank you for this Rich panel um looking at the title we heard a lot of sophisticated challenges of monetary policy but I would like to take you again to the poly crisis New Normal having representatives from three policy institutions what makes you lie awake at night in terms of Str structural shift and or crisis which may have the most damaging effect on inflation and I mean with damaging effect inflation on inflation a deviation upside or downside from the Target and or volatility what is it what really because I mean if I had asked you these questions in 2019 nobody would have said pandemic and it’s a very personal guess but I really would like to know where you see the dangers looming is there any other question yes please France thank you uh just recently in the eur50 group I had a long discussion with otma ising uh and he said the original monetary framework of of a eurosystem uh crosschecking with monetary Aggregates uh something like atasia has proposed and we have had it would have been much better when current framework uh it’s my question to Frank what do you say okay so we have two questions what keeps you awake at night uh in terms of inflation scare so maybe we all all three of you can can answer that small survey totally representative and the and the other question is uh should we have returned to monetary targeting I mean to to paraphrase it and exaggerate all my think he said crosschecking okay pillar two okay so let’s again start with with the final round with with fella yeah so I I think that uh several factors can can can uh can several of the structural developments can can be equally dangerous I would say that the the the supply Factor simply because they’re very tangible at the moment with the energy transition with the geopolitical tensions which seems to loom large in the future um are those that present a biger risk to me to the inflation Outlook at the moment but the others particularly our star might be less tangible but depending also on the misperception could also be potential contributors um and uh and of course the crosschecking is part of this broad approach to robustness of monetary policy so I think it’s very important thank you Frank what keeps you up at night I mean many of the issues we we talked about uh but I mean I must say much less from the perspective of a central banker and a monetary policy advisor than just from the societal challenges know that climate change I think these are so or another way of putting it I do think our current framew works are appropriate I mean we will sort of know make changes and and review our current strategy uh by 2025 um but uh in the big scheme of things I think we know how to maintain price stability again there will be hiccups but also most recent quite extreme example shows that we can uh deal with that so so when I’m I’m I’m I’m awake at night I I worry more about no the other implications for welfare uh of no geopolitical risks in particular but not so much about uh that this complicates our life as Central Bankers I think we can we can deal with that through these robust strategies uh and even if we make mistakes I mean we have powerful tools I think that’s one of the lessons we’ve learned to to sort of catch up and and and keep uh price stability uh in check um I mean on on the monetary uh so so one of things we did in the last strategy review is actually to integrate no we we actually had had a look integrate our monetary and financial analysis with the real uh analysis and and so and that’s very much in the in in the sense of of of crosschecking um but the reason why we did that is exactly because it’s very difficult to interpret what’s going on in money growth and credit growth data without know understanding also what’s going on in the real economy and some of the other Financial uh markets so what that’s one of the things that would this particular frame our old framework have prevented I don’t think so at all I I I I I don’t think so at all I think the biggest uh shock came from something that is completely outside of of of our ability to control which is a war in in in in Europe of course we have to deal with with the consequences I don’t think changes in the monetary policies would have made a big difference and and there is quite a bit of no of of also actually Source simulations I think uh that uh the Bas has done that where you look at alternative policy responses yes you have some effect but the big picture doesn’t really doesn’t really change thank you Mikel what keeps you awake at night and what about the second pillar um well so I I think the um my my biggest worries are um linked to more structural things sort of longer term developments whether it is about Slovakia or the uh or the Euro area but so that has uh um unclear consequences for inflationary development but when it comes to the inflation itself I think the uh the biggest worry is about uh um the the fragility of the of the world there a Slovakia is a small open economy a part of a uh a large open economy um that um basically you know we have a we have a war going on next door which uh for Slovakia has important implications also in terms of energy Supply uh we um we we sort of uh storages are full uh the narrative is that the looking to the Future there is diversification um going on so things should be uh uh safer U at the Euro area level it’s probably even better than in the context of uh just Slovakia but all that new world is uh at least to a certain for the foreseeable future is meant to rely on a on on a Global Network of uh sort of LG and and and these kinds of supplies which is inherently more fragile than uh the kinds of supplies that we had so uh you know you can get um these energy shocks uh unexpectedly um uh quite a bit obviously there are other u i mean the whole sort of societies are are more fragile and more fragmented uh these days which uh which may have implications um which brings me to the second B and that’s about fiscal policy um I have a past in that area and and it seems to me that somehow the uh uh the in in a lot of countries in a lot of places you barely see a force uh that would be currently advocating U fiscal responsib so it has kind of gone out of the uh political debate completely and it’s an important precondition for um central banks to do that uh to do that job whether it’s going to be the market who will uh bring the topic uh back into game hopefully not um ideally you would want to build a kind of new consensus around uh some kind of U fiscal Prudence um going ahead because there are big challenges and we haven’t still sorted out the uh the past uh from uh in the area of fiscal policy thank you very much uh to this excellent uh round um I think we learned a lot from from you three thanks so much for sharing your knowledge I hope be we we satisfied your uh your question in a way I take along that basically the panel thinks that on the monetary policy side on you know as far as the technicalities of monetary policy I mean we basically know what to do but uh the real problems come from shocks political geopolitical supply side factors energy was mentioned societal developments that are Beyond fiscal policy that are basically beyond our control and uh that there for which there are no easy answers uh you know for monetary policy okay I I think this was really interesting thank you so much uh [Applause] thanks okay um thank you very much for being at the last sessions so this is an academic session about inflation new insights so we have three great presenters um first of all we have rre Garcia from the ECB and he going to present the return of inflation on inflation risk so one you have 15 minutes okay thanks very much first to the organizers I mean for giving us the chance to present this work here I mean it has been an excellent conference and uh this is join work with Ricardo gimeno Pierce H and Michael I mean Ricardo jimeno is from the bank of Spain so the usual disclaimer applies not only to me and the ECB but also to the bank of Spain and the Euros system H Pierce uh how Michaela from the University of Noland in Department of mathematical Finance so uh let me skip this I mean the motivation for for this paper I mean is is fairly simple I mean we have been discussing this I mean this over these two weeks and from for long already I mean inflation has fluctuated significantly off Target in recent years in particularly in the Euro area both in the past decade below Target and then we have the the post-pandemic inflation search and in this context inflation expectations are uh closely monitored the problem is that those most indicators of inflation expectations I mean tend to provide only the central scenario H and we know that monetary policy decisions in particular but economic decisions in general I mean will benefit from having probabilities assigned to the different alternative scenarios beyond that that Central scenarios so in particular I mean um uh one of the goals of this paper is really showing that evidence on inflation risks across different Horizons I mean can help better assess the Persistence of inflationary shocks and also the challenges of price stability over the medium term so just to put those comments in context I mean h the two charts uh you have there I mean they show for the Euro area and the US developments in a year on-ear inflation rates the solid black line what you see I mean those those fluctuations but also indicators of long-term inflation expectations that we consider normally I mean quite quite stable in particular I mean the dots you see in both charts are survey indicators while the blue lines are the standard measures for long-term inflation compensation measures the five year in five year uh capturing inflation expectations between six and 10 years ahead I mean the point of this I mean we have been discussing I mean the the uations I mean in inflation H during this conference but also the inflation expectations I mean has fluctuated significantly for example in the Euro area from around 2.5% uh pre Global financial crisis I mean we reach a bottom of around 1% um er in the after the covid outbreak and then a very strong rebound to the pre Global financial crisis levels I mean with inflation surge again as I was pointing out I mean this shows only the central scenario I mean the goal of this paper is really to explore H inflation risks surrounded those Central scenario for inflation expectations at different Horizons and just to be clear I mean we obtain those um inflation risks from using inflation options traded in both the Euro area and the US market so in this paper we remain with the the risk neutral um measure we don’t make any attempt to disentangle the inflation risk Premier component I will come back to this later in the presentation so this paper makes two contributions I mean to this uh to the literature using inflation options first of all I mean we introduce a robust methodology for estimating this inflation R neutal densities for robustness H I mean introducing the minimum set of assumptions allowing us to really exract I mean this time structure of inflation risks in particular I will show you evidence for inflation risks at the 2-year Horizon shortterm the medium-term the fiveyear and the longest term forward five year in five year that matches the standard uh metric for the inflation compensation in terms of interpretation I mean we look at the inflation risks and then document I mean uh the evolution of inflation risk both in the Euro area and the us since 2009 which is when we get access to these uh these inflation options uh being traded and as a as a sort of corollary of our analysis I mean we try to assess the risk for Price stability posed by the inflation surge in the post-pandemic period but putting this in context to the historical evidence we have from the previous decade that in the case of both the Euro area and and the US I think it’s important important to understand I mean the current situation uh I’m not going to to elaborate I mean on the on the related literature but this is part of the limited but growing literature using this inflation options Market I mean this my personal interpretation of this this literature is has evolved from the original work of kidan right just only focusing on one horizon in that case I mean the five year uh two attempts to really estimate I mean with different methodologies the horizons of interest in particular the longterm forward h r neutal density so uh before showing you the results I mean let me just give you I mean a brief overview of the key in um messages I want to make uh honoring I mean the title of the of the H this session I hope to to show you that inflation risk neutral density this provide important insights and these insights I mean go from allowing monitoring risks at different Horizons assessing the Persistence of inflationary shocks and also some Noel information on expectations formation uh because we move beyond that point forecast into a density forecast and the Dynamics of that metric I mean are much richer in terms of interpretation I mean we document that there are significant differences between the Euro area and the US inflation risks in particular I mean the FED that the ECB have faced very distinct challenges uh to price stability I mean in the last in the last decade and turning into the common shock that has was documented by by um Fel I mean was quite symmetric across many economies H the recent inflation search the challenges to price stability seems to be relative ly contained I mean on both sides of the Atlantic I mean a bit more in the Euro area where there are some potential H vulnerabilities I mean in the formation of us expectations and in any case I mean this uh despite that being relatively well contained I mean inflation risks I mean deserve close monitoring and the type of indicators I mean I’m going to show you these inflation R neutal densities could be very instrumental in that regard uh I’m not going to describe in detail I mean the methodology I mean unfortunately I mean for this I mean the devil is really in the details but I’ll give you a big overview of what the main issues are uh we have a strike prices for inflation caps and Floors these caps and floors for inflation I mean work as for any other Financial instrument like interest rates or or FX in theory they are available of a large number of strikes up to 14 but in practice I mean liquidity and Market activity is concentrated in very few number of of both caps and Floors uh on average I mean for the since 2009 I mean I would say you have access to four caps and four floors ER active over over the whole sample that imposes some some uh restrictions in the estimation of the RIS neutal densities and that’s what I refer to as a robust methodology can handle I mean these limitations of the of the data but you have to clean the data uh significantly H we also use inflation uh link swaps and ois rates I mean the derivation of this uh of this uh rutal densities and it’s important that I will talk about two different kinds of R neutral densities first the spot rutal densities are estimated directly for the traded Horizons for example the year the five year or the 10 year for that we use a nonparametric approach that without imposing any sort of functional form allows us to to accommodate any potential asymmetry of fat tales that are in the data H and in terms of practicalities I mean we we H Implement plan interpolation and extrapolation in the volatility space as has been shown in the literature that is the most way imposing no Arbitrage conditions uh the second are the forward RIS neutral densities that are needed to be derived for the non-traded horizons of Interest like for example the 5e 5e for that we we propose an student copula that provides I mean significant flexibility over the 16 methods in the literature so let me show you I mean the the evidence on the RIS neutral densities I mean this is for the Eur area shortterm medium-term longterm I mean the key message for these uh charts I mean is that there is this predominance of these blue uh lines that reflect the probability assigned to inflation outcomes below the 2% H level and this holds for most of the sample add in the recent inflation surge not only at shortterm Horizons and medium-term but also and long-term ones a more stand representation I mean of these probabilities is in terms of the upper tail for high inflation or lower tail H you have the three Horizons there the darker the the line the longer the Horizon but I would like to draw your attention for interpretation in a way of summarizing this to the balance of risks which is calculated as the probability above 2% minus the probability below 2% we found this indicator particularly interesting because it provides not only a Direction signal on the risk to price stability whether you are likely to miss the target from above or From Below but also of the magnitude of that risk given the difference for example in the case of the Euro area I mean you see that short and medium-term um balance of risks I mean decline significantly into negative territory early on in the past decade and only rebounded strongly with the with the um inflation surge uh post pandemic in importantly in the Euro area you also get I mean a decline after holding for a while I mean also the long-term uh risks I mean dive into negative territory H suggesting I mean there was a significant entrenchment of low inflation risks in the Euro area in contrast in the US I mean despite some episodes I mean you see that there’s a much more balanced particular medium-term Horizons and in particular a long-term Horizons I mean between above and below 2% this is reflected in the Dynamics of these balance of risks I mean the dark line I mean as you see I mean fluctuates over the whole sample is slightly above the zero level I mean pointing out to predominantly upside risks I mean to inflation in the US in contrast to what happens in the Eur area Beyond this documentation of this magnitude of the inflation risks I mean we use a simple pass through model as the one Vincenzo was was introducing I mean this this morning to analyze the Dynamics of these inflation risks I mean and then I would show you evidence on the pass through from backward looking information basically actual inflation developments but also on forward looking information embodied in the short-term inflation expectations how this has translated over time into medium and long-term inflation risks uh turning to inflation I mean uh there are some episodes the the blue uh charts are for the Euro area the red ones are for the US and the top panel is medium-term the bottom panel long end H there are some episodes I mean of of a stor pass through I mean from actual inflation but particularly at the long end I mean there’s very little action even within the inflation H search uh so this is we believe I mean this is part of the problem of uh mechanically inferring that actual inflation developments are going to be translated into the pricing of these of these risks we do a more structural interpretation of the investigation of the inflation Dynamics decomposing them between Trend inflation capturing persistent influences and cyclical influences I mean that chart I mean shows this DEC composition between these two components and when we reapply the the pass through there is a very strong and significant pass through at medium term but despite the fact that for the Euro area in that late last decade H significant period it has significant pass through it has declined gradually the inflation search the only thing it did is seems to attenuate I mean they declin but they has declined now into nonsignificant territory same analysis for the US I mean points to the very strong impact at medium-term Horizons but also confined there and not translated into long-term inflation risks we interpret this together with the evidence from the using overall inflation I mean we interpret this this evidence as suggesting that this inflation risk rice not only inflation at face value as observed but the markets really take into account the persistent components that could be important I mean for these medium and long-term Horizons in their in their pricing in terms of the likelihood of the entrenchment of inflation risks I mean we look at the pass through for some term uh inflation risk into medium-term and long term as before blue lines for the for the Euro area red lines for for the US the top panel is the medium term that is strong and uh significant over the whole sample pass through at medium-term Horizons but there are important differences in the long-term Horizons in the Euro area it picked at around uh 2018 at a significant level and as I pointed out before I mean has been declining thereafter to become I mean nonstatistically significant in the US instead I mean it has been relatively muted in magnitude but has remained statistically significant for most of the sample what makes the long-term inflation risks in the US more vulnerable to sifts in these short-term risks I mean should the the inflation ER U or the last stage of the convergence towards Target I mean become um a bit bumpy uh those are rs but this is I mean just let me conclude I mean by some uh remarks I mean highlighting I mean the key messages I mean uh those inflation RIS neutal densities ofer important insights I mean for the inflation Outlook and in particular I mean we can monitor using them short medium and long-term uh Horizons H in terms of the the evidence I mean both the US and the Euro a inflation risks they have completely different characteristics in particular I mean over the past decade that should also be taken into account to interpret I mean the situation right now where the pass through from certain expectations and risks suggest that risk to price stability are already relatively contained but close monitoring of these indicators I mean could provide I mean important insights I mean in the last stage of convergence to to Target thank you so so we have time for a few questions so is there any questions fting presentation uh for me the question is you have a number of interesting charts there and the end you came out well us and Europe I to say has broadly the same effects there but before you had different charts there I wasn’t aware whether the charts allowed to explain why the inflation persistence in US is higher than what we was we see in Europe does this come out of Europe estimates there I mean I see different developments but for me what I actually observe is that the US the inflation is stickier than in Europe well my question is do I seeed some in their child in order to explain it this would be for me helpful uh in the chge if I can go back at least you didn’t allude to it what for me well for example the the pass through to shortterm and medium term from from actual inflation is much stronger in the US that it is in the Euro area that that you can see I mean for the comparison of the panels so this would be an indication not only not only sorry only something which one see and you would Cally link it that this higher path through is one of the reasons why inflation is more persistent or stickier than in in the in Europe this would be your explanation okay I was missing this in your presentation because this is something which is important for a policy maker thank you okay I understand think we have a second question here thank you very interesting uh I did similar work a few years ago um so I was wondering as so you talk about inflation risk back then we also distinguish between the inflation expectation and the inflation risk premium which can also be inferred from market prices did you also are you also looking into do that so the inflation risk premium um that’s that’s in the agenda but first I mean we wanted really to understand I mean the Dynamics of these risk neutral measures I mean before embarking into into this I don’t know at to which Horizon I mean you use this DEC composition but that’s that’s one of the of the remarks I mean or suggestion for f work I wanted to flag I mean didn’t have time with the presentation but that’s there I mean the problem is how you do it I mean you need to impose very strong assumptions I mean to decompose that H normally the approach has been either you derive a physical density out of the micro model or a Time serious model H but then you then when you contrast this to the market that may not be I mean the one the market is pricing so we think I mean the best uh the best way for I me would be to join estimation of both the physical and the RIS neutral densities then you can use the power of the time series and the macro models at the short end and the power of of the information from the market at the long end yeah thank you I think that would be very interesting to see yeah thank you definitely uh we have time for one or two more questions are there more questions I mean I have a question if you if you think about your paper are you really looking at inflation risk or are you looking at the perception of risk from the financial markets yeah I mean when you remain I mean in this because if you think about it when you look at I’m not so familiar with the data set but when we look at the SPF it’s kind of pretty funny because they make quite a lot of mistakes and their forecast are not that great and even when inflation is very high they still make Auto qualed mistakes so do you really think you’re looking at risk or you think at some kind of assessment of the market expectations well I mean it’s like when we we look at inflation compensation I me we talk about inflation expectations of the private sector I mean this is the private sector assessment of inflation risks in each economic area by definition so this is how you would spin okay yeah I use inflation risks for short but that’s what I stress that okay we remain with the r neutral the r neutral measure perfect is there any question online or any more questions okay so thank you very much so next next we have Jean Paul ler from BR University and he’s going to present a more theoretical paper which is called can supply shocks be inflationary with a flat Philips SC great so thank you very much uh it’s a pleasure to be in such a great conference I I I had a lot of fun and learned a lot so uh like Alex said this is a little bit more of a theoretical work but it as you will see the lessons that come out of this this approach for the implementation of optimum monetary policy are I think quite thought provoking and so you know in the spirit of the conference I’m going to try to emphasize that uh throughout the presentation so this is called can supply shocks be inflationary with a flat Philips curve and it’s join work with Greg Fallon so let me set the stage by laying down two facts about business cycle data in particular with regards to inflation fact number one is the Philips curve is very flat so what we mean here is think think about the housing bubble of the early 2000s think about the Great Recession think about QE 1 2 3 and 4 those were episodes where the output Gap was very large either positive or negative it doesn’t matter or there was a large amount of monetary easing and despite those developments inflation remained remarkably stable okay given these episodes there’s now a growing literature looking empirically this the slope of the Philips curve so papers by you know the Negro and others Hazel and others very important work looking at the slope of the Philips C fact number two is well Supply shocks are inflationary think about the 1970s think about the post episode postco episode that we just went through inflation just moves a lot when Supply shocks hit and so the starting point of the paper is going to be the following we we’re going to point out that despite these two facts in our opinion being very important to characterize inflation Dynamics these are seems to me like seems to us like first order facts that that we want to f with our models it turn turns out that standard models cannot fit these two facts at the same time so why is that if you have a model that gives you a flat Philips curve then it means that your price your price is very rigid your price level is very rigid okay but then a very rigid price level cannot generate inflation cannot cannot give you inflation out of Supply shocks and this is a shortcoming of the calvo it’s a general shortcoming of standard price inflations Calo Taylor Rottenberg and menu call so I’m going to develop this point a little bit more in the context of the newel model later on okay and so why are we going to propose in this paper giv given that shortcoming what we’re going to say is that the data actually want a model that has the following two features feature number one is prices are sticky when demand shifts that’s going to give you fact number one of the flat Philips curve but at the same time prices are flexible when when Supply shifts and so the conjunction of these two features we’re going to call shock dependence and so to make it clear so this is different from the important Notions in the literature that have been considered as time dependence State dependence so State dependence for example is different from State dependence State dependence is what matters for whether firms adjust prices or not is how far away they are from their optimal price so this is different what we’re going to say is we’re going to maintain the the the we’re going to the approach here is that it is the type of shock that hits Farms that has implications to whether Price Farms are going to react in their pricing decision or not and so the name of the game is going to be given that idea is going to be to provide a micro foundation for a shock dependent price INF friction this micro Foundation we’re going to provide is going to be based on a strategic interaction between firms and consumers and that strategic interaction is going to lead to two results firms are going to avoid increasing prices when demand increases so that’s going to give you fact number one but at the same time firms are going to pass on cost increases to Consumers without any problems so kind of to give you an idea of the type of behavior firm Behavior I have in mind here let me show you this uh empirical evidence so the paper has a very short empirical section which is this one uh and this is obviously subject to interpretation right but what we have in mind here is basically the idea that we find fascinating that firms do want to justify increas in prices as a need to cover increases in costs so for example this is from Anna sakaria that you know if you have been to the Boston area for conferences or for other reasons which is is a Mexican restaurant in the Boston area that has great uh great food by the way so you should go if you go says our they posted this sign actually back when I was a a PhD student at MIT and I actually find found it so fascinating that I grabbed the poster the sign that had posted and asked him if I could make a photocopy of it and this is what what what I what remained of it they said our prices have changed due to an increase in costs we have raised our prices we are proud of our low prices and make every effort to keep our cost down thank you for your understanding and so to clarify what we’re going to do here is we’re going to have a model that we’re not going to have behavioral assumptions so we’re going to have a model where all agents are rational but this type of firm behavior is going to emerge out of our setup due to the Strategic environment that we’re going to write down okay and so why should we care about this well in this this conference where we have kind of talked about so many fascinating facets of of monetary policy and more more generally Economic Policy these are the Aggregate and policy implications so supply shop shocks despite you being able to have you know long periods of time where inflation just seems to be dead Supply shocks May inflation come alive and in fact there is no price dispersion in our model because prices are flexible when Supply shocks hit meaning that the reaction of prices when Supply shocks hits is efficient so there’s a very HCK hakan uh kind of type of flavor in our paper Okay so Fred haek you know he said well you know you should let the price system react to what’s going on in the economy and that’s exactly what happens in our model when cost increase because there has been a supply shocks firms increase prices and so there’s no business for the government intervening and and dampening the the increas in prices because it’s actually the efficient reaction of the price of the of the Goods Market okay so the the the inflation that you get out of Supply shocks is efficient and that’s something I’m going to stress and I’m going to show some simulations down the road if a central bank raises rates because the central bank doesn’t like the raising inflation the central bank is going to generate a negative demand shock and remember so in our model we’re going to have a flat Philips curve at the same time meaning that the sacrifice ratio is going to be very big potentially infinite so basically with a flat Philips curve there’s little or no effect in inflation so the the the action of racing rates in order to bring down inflation is going to be completely ineffective and the central bank is going to lead to is going to generate a big demand uh um uh uh uh fall a fall in aggregate demand that leads to a welfare loss and the reason is the demand shock is inefficient okay and so also you know when Supply conditions normalize inflation can come back down seamlessly when Supply because you know firms have flexible prices to supply and so when Supply conditions react the inflation episode is over so those are the policy implications that the paper provides and I’m going to go back to them uh in a little bit so let me go let me talk about this shortcoming of the standard model so let me look at the supply shocks in the nuian model the equation that you see on the slide is very familiar to all of us that’s the nuian Phillips curve in its structural form and what I want you to notice is the terine red is the markup shock right in the close economy version of the model and it has a coefficient Lambda in front of it Lambda and Kappa are the slope of a Philips curve okay and this is a structural Philips curve now this Lambda is estimated by a wide range of papers in the literature to be very small so for example the estimate that I’m putting on the on the slide 020 comes from the paper by the nego primeri lensa and talotti on this on the on the Philips C so what does that mean so the way the the literature proceeds is a normalization of this markup shock calling it sort of a a shock another sort of reduced form shock new which is Lambda Z which means that you know to us this is a little bit of a tricky normal ization because what you’re saying is if I want to generate a 1 percentage Point increase in inflation then I need a shock a markup shock a microfounded shock that is 1 divided by 020 meaning of a size 500% and you know as all of you as know when you have Supply shocks episodes inflation increases by more than 1 percentage point so if your Set Side markup is 12.5% which is kind of a standard value that I took from you know G textbook value I took here from gal this means that the new desired markup of firms has increased to 575 per. okay A number that seems questionable to say the least right and so what’s going on here so what’s going on here this is actually so you know the calvo model is a great model we all like it because it has very nice properties is very tractable and for that for those right reasons it has become a benchmark model for the analysis of monetary policy but what we want to highlight here is that when it comes to the analysis of Supply shocks it is problematic it is not it has a feature that is problematic and and the feature is that the the model implies the same degree of stickiness for all shocks and so if you want to generate a flat FIP curve then you end up with this type of implications for for Supply shocks so this is related to the wellknown to the of to some of to some of us that work on this literature charatan critique but here what we’re emphasizing is that this is intimately linked to the be price Behavior implied by the calba model okay and so you know I mean this is kind of relatedly for example the literature has has has tried to solve this problem for example Frank has a paper in 2006 where he proposes to think about markup shocks as coming from a flexible price sector this is a paper that he wrote when he was writing in par parallel I guess he’s very influential War convesion analysis of of business Cycles using DSG models he proposes to analyze to think about markup shocks as coming from um a flexible price sector so what we’re going to do here is slightly different we’re going to say is perhaps not a flexible price sector but is perhaps that all sectors in the economy have this Behavior where prices are more flexible when Supply shocks hit okay and that is going to carry important policy lessons so why don’t I go ahead and tell you what the type of model that we have in mind so and I’m going to do this providing some intuition first I have five minutes left so I don’t know how much of the equations I’m actually going to be able to provide but you know I I’ll give you I’ll try to give you a flavor so the the key assumption is that we have superiorly informed firms so basically in our model firms have more information about aggregate shocks than consumers so imagine you know firms looking at the wage Bill and having more information about the supply side of the economy or looking at the de kind of total sales of the past few days and figuring out faster than consumers what aggregate demand is and so the crucial point is that that Assumption of firms having better information together with price setting is going to imply a strategic interaction with consumers that is going to lead to the following results so Supply shocks are going to lead to price flexibility and demand shocks are going lead to price stickiness why is that um let me go what step by step and the supply shocks The crucial point is that costs are not pay of relevant to Consumers consumers all they care about is the parameters of the utility function their income in a decentralized market they just compute their demand firms look at the at the demand they set a price to maximize profits and so you end up with your very standard micro kind of inter like undergrad micro Monopoly problem where obviously the price reacts to the marginal cost of the farm and there’s actually no strategic concerns between firms and consumers prices are flexible things are more interesting with demand shocks because now information that the firm has is pay of relevant to the consumer so basically uh let me tell the story in the following way here what happened is that the firm okay so the information about aggregate demand is pay of relevant to the consumer what mean meaning that in a model like us where there’s aggregate demand complementarities if I know that aggregated demand is higher that also means that I should consume more and should be willing to pay kind of a higher prices because my demand should shift outwards but firms the problem is that firms have an incentive to misrepresent the state because the firm makes higher profits when aggregate demand is high and so the firm wants to basically pretend that aggregate demand is high increase prices in in order for consumers to spend more but rational consumers understand this and therefore they do not they price increases are not necessarily credible in equilibrium therefore there’s a strategic friction because of this incentive problem of the firm and that leads to price thiness so kind of a a way that I like to tell the story is let me tell you a story is you know what I call the like movie goinging problem yes I know let me just tell the story because I think it should be clear so the movie movie movie going problem is you know it’s like when a friend tells you friend wants to hang out and tells you let’s go do this and I know that friend your friend likes doing that and so you’re a little bit suspicious if you should do that right he’s like I have the problem with my wife all the time my wife loves watching movies and so for her like going to the cinema is great all the time I don’t mind going to the cinema but only when movies are really great and so you know my wife is going to tell me you know why don’t we go to the movies tonight I think I’ve looked at the reviews of this movie and it’s great I think you’re going to like it and then I’m I’m thinking well maybe I I don’t know if I should trust that because not because I don’t trust my wife but it’s because you know she has an incentive here which is to go to the movies here’s the same The Firm has an incentive to raise prices because the firm likes that and so consumers are a little bit suspicious when they are trying the firm knows what the aggregate demand is so that’s exactly what happens and with cost shocks this incentive problem doesn’t arise because Rising raising prices when cost increase is credible because the firm cannot affect the the demand of the consumers anyway so I clearly what the how long how how much I’m I’m I’m going to skip the details of the model what I want to show you is basically what I get at the end of the day so basically this proposition gives you the Philips curve that we find so this Philips curve says so there’s a cut off in the fraction of informed consumer so that cut of uh uh uh governs how much M how much information a symmetry there is between the firm and the consumer that’s not important now what’s important is the equation you see on the slide so the Philips Square we obtain is inflation is equal to Capa times the output Gap plus the supply shock so notice that there’s no coefficient in front of it this is kind of your old style Phillips curve that shifts up and down with Supply shocks and the point we’re making is the Newan model doesn’t give you that and so basically a way to think about the paper is that it is okay to kind of the practice of appending a a markup shock to the Phillips curve is correct is a good practice but basically what we’re saying is that implicitly when you do that you are assuming that prices are more flexible when Supply shocks and so you know our paper has done the math to obtain that type of that type of equation is basically you when you do that at practice of appending a a markup shock to the Philips curve you are already assuming that prices are more flexible to supply shocks so that there is shock dependenc in a sense and so now inflation can move one to one with a supply shocks like the the way we liked in like kind of like old style macro okay so these are the simulations I promised these are the effects of the supply shock so that is plotting inflation output Gap interest rate and Welfare okay and the colors are different hawkishness of the Central Bank the blue is a central bank that doesn’t react to inflation at all red is a kind of a Midway hawkish Central Bank a coefficient on the interest rate rule of 1.5 and three is a more hawky Central Bank of with a coefficient of three on inflation in the tailor Rule and so as you can see all C Banks virtually achieve the same outcome for inflation that is because the Philips curve is flat in our model so basically raising the interest rate and affecting demand has no effects on the output Gap because firms do not react to to to their prices do not react to demand okay so we get that’s what you get with the flat ctive curve but the more hawkish Central Bank generates a negative output Gap the intermediate red generates a Midway output Gap and the blue one doesn’t have react doesn’t generate an output Gap at all that’s the efficient reaction uh of the Goods Market so output has gone down if I would were to plot output you would see that output goes down but there’s no output Gap Supply shocks don’t generate output gaps in our model the reason is that prices are flexible so inflation increases and uh uh demand adjusts but there’s no output Gap okay and you know obviously you can see that the more Hockey Central Bank is the one that generates the biggest welfare losses the central bank that doesn’t react to inflation is the one doesn’t generate any welfare losses at all and uh so the takeaway is idea this idea of shock dependence and how it changes our our way to thinking thinking about the role of monetary policy in the face of Supply shocks okay just to clarify this is the the model that we have in the paper is just one model that gives you shock dependence but there’s obviously other models you could think about and we think that’s kind of an interesting Way Forward when it comes when we’re trying at least empirically to capture the effect of Supply shocks that we all know tend to move inflation a lot whereas demand shocks do not thank you very much so thank you je Paul very much for the presentation is there any questions um uh I think yeah sorry you’re the next one okay uh Daniel Hardy from the Viet Institute for international economic studies I was wondering how you might think might empirically test this versus an alternative where just some sectors more flexible than others and you know you have commodity price shocks that against transmitted to oil prices and that was drives it versus your model okay uh can I have the the last of the slides again so we have a test that I obviously didn’t have the chance to present which is just simple V uh previous slide please oh actually I can move that previous slide this is showing the reaction of the US CPI to demand versus supply versus demand shocks so for brevity let me not not tell you how I did this but this is with external instruments that I that we grab from the literature and so for a normalized effect of industrial production at 24 Horizons onwards um you know the the the Orange is the demand shock so it leads to a positive effect on industrial production the the blue one is is is Supply it leads to a negative effect on industrial production both raise inflation you can see that the US CPI reacts by more in the case of Supply shock so that is consistent with our model I’m working on a paper where we have survey evidence where we ask firms why do you increase prices for example we have a survey of German businesses at the opening of the of the of the economy in March 2021 after covid where many of those businesses that we have in the survey increase prices by a large amount and we asked them why did you increase your prices did you increase your prices because there’s pent up demand because you have a line because you you’re you’re not these are hairdressing businesses actually so you know they we we see their booking systems like they cannot accommodate everybody there’s a line and they literally very clearly say we’re not reacting to demand we’re not trying to close demand what we’re doing is we’re increasing prices because of the hygiene costs masks uh um this disinfectants all of distancing all of the procedures that that were in place after covid that increase their cost and they’re increasing they they they they say they’re increasing prices due to that um you we’re running a survey where we’re going to ask consumers do you think it’s okay for firms to increase prices when their costs have increased so the other side of the coin um another better even better way that I haven’t uh develop very much in my AG in my research is uh looking at the Micro Data and looking at you know when if you can could if you for identified uh supply and demand shocks which of those generate more price flexibility in the Micro Data but you know that’s what I can tell you so far okay thank you uh yeah there’s a short question uh to understand your counterintuitive results so if I understand correctly in your model if there is strong demand uh firms simply increase the output right so it’s not the case that there’s excess demand that cannot be satisfied and it’s not the case that firms increase their price to bring up a new equilibrium and but it’s really the case that even though firms uh have a supply shock perhaps they still increase output uh in order to satisfy this high demand without increasing prices which yeah so that’s that’s what happens so it depends how you define excess demand you know the haircut hair cutting businesses survey there is clearly a CL number of clients that cannot be served but they’re going to be served down the line so it depends how you think about the the functioning of the market but in our model yes so Farms the problem is Farms cannot credibly increase price and that because they have an incentive to do so in in all states of the world so there’s this strategic friction and so in equilibrium the pre the the outcome that arises is sticky prices which is a a firm optimal is a optimal way to deal with the incentive problem and therefore that means when demand goes up firms are going to sell more but they’re going to keep prices unchange when demand goes down they’re going to keep the same prices we’re going to they’re going to sell less that’s that’s what arises in our environment and I’m happy to kind of try to explain the intuition again or go you through the go through the equations of the model to explain how that works but the B the the basic point is the firm wants to increase prices The Firm likes it and so if I can convince you then I would make higher profits but therefore you would not believe me if I increase prices and so therefore in equilibrium what I end up doing is not increasing prices at all my prices do not move with aggregate demand instead when my cost increase I tell you hey sorry my cost have increased uh you know that’s what it is I need to keep my business going and so I’m increasing my prices and that’s credible because you know the firm wouldn’t do that if the cost had gone down if cost go down the firm lowers prices that’s what’s good for business and so that’s why there’s no strategic friction in the case of costs that’s kind of like the Insight that in my opinion the paper provides in the terms of in terms of the behavior of of farms that are that have more information than consumers one more question y fik my question is uh wouldn’t it be the more efficient reaction of a central bank for example if you have an energy price shock to uh Finance additional energy Supply when you have to increase prices because when you react to the supply shock and uh not to with demand chock and in your paper in your uh environment this would be more efficient that would be welfare increasing I it would not be so it would not be efficient in the sense that you know if there’s a shock of course if you think you can deal with the shock is even better I fully agree with you you know is like the Press was saying you know why is why are rates is why is the Central Bank increasing rates it’s not that that’s going to bring more oil to the economy right so you would expect you know the best thing to do would be to deal with the supply shock directly in order to avoid the increasing prices I fully agree with that what the paper is just saying is increasing rates it doesn’t bring down inflation in in a in a model in in this type of economy where a whether you have a flat scurve and it it generates a welfare loss that’s what the model is saying I fully agree with you I’m I’m 100% on board with that okay thank you very much this is I guess a very provocative [Music] paper so next now we have Ken kber who is my colleague and she’s going to present maximally forward-looking Co inflation uh the floor is yours thanks uh thanks a lot for the short introduction I have the honor to give the last presentation for this conference and it will be about the maximally forward looking core inflation uh this work is co- offered with um Philip Gully colum and Christoph Brett from the University du Quebec Montreal and also with maximilan Gober from bonei University since I’m affiliated with the Austrian Central Bank the usual disclaimer applies and everything I say and I show you today um are our own views I would like to start uh by showing you inflation developments for the EUR area so here you see the hicp inflation rate um in year-over-year changes which we usually refer to as headline inflation you also see core inflation so this is hicp excluding energy and food and the trim means and uh what we see from this plot here is that first of all the two measures so core inflation and trim mean that are supposed to kind of show us the underlying uh developments in inflation are less volatile which is by uh construction uh but also that they are often lagging behind a bit so especially when you look at the recent search uh for example we see that core inflation but also the trim mean um they lag behind and this uh also holds for other countries so for example for us us I mean for the US it holds that especially for the recent search um so let me mention here you see the PC inflation rate and core PC and the TR mean and for the recent search um for the US I mean we know that in the US uh the search was mainly driven or a key driver uh was Goods price pressures right whereas in the Euro area we really had a a large commodity price shock this is why the core inflation rate is here more timely but still you see that the trimp mean is uh like behind and this still holds when we look at more timely Transformations so here for example instead of showing you the year of year rate which we usually look at uh we can also look at the quarter over quarter rate and uh again for the trim mean you still see uh this is kind of lagging behavior and before I get to our contribution um let me show you or let me give you a quick overview on existing core inflation measures and how they are built and also the the properties we would like them to have so the the measures I showed you uh so far they are built on horis sixs so we have permanent exclusion measures um like the core inflation rate so a prior we decide to exclude uh energy and food um there are extensions so for example in the US uh it is um also common to look at PC super core which is then core Services minus housing uh we also have temporary exclusion measures like for example uh the trim mean or median um CPI where you exclude um the tales of the distribution so you take the monthly price growth distribution and you exclude the Tails and then we have modelbased measures and this is um certainly not a complete list there are many measures out there but I would highlight uh two very important ones that we are also tracking um and following um this is the multivar core Trend um me from stock and Watson from the FED um New York here and it’s the ccci the persistent and common component of inflation uh that is reported by the ECB and both measures here they are based on unsupervised techniques so it’s a factor model approach that is um extracting the common component um of the underlying data so the underlying a price growth series and um an important question is uh what do we actually want to have so what are the properties that that we would like to get and first of all this is low bias so this means we want to get the same or similar long run average than headline the second is low volatility so in a sense we want to extract signal from noise and the third one is especially when we are trying to pursue a forward-looking monetary policy we’ also like to have a low error when we forecast headline inflation so then we can really see uh where where inflation is going in in the medium term uh the last Point uh would be also nice to have this is rather subject to another project um that we are doing so it would be nice if it was also highly reactive to monetary policy and um reluctivity conditions because what we then would see is really the parts uh that we can actually influence as as central bank right um so our contribution is the following we are really focusing on this third um third property that I showed you so we really want to get a good forecasting performance and we do this in a supervised manner so we write a very simple U machine learning um regression problem as we take a simple um machine learning algorithm that kind of focuses on getting good predictions um so we’re trying to so in fact we’re trying to assemble and aggregate inflation subcomponents that so that the aggregate is maximally predictive of future headline inflation and uh we call our model the or this regression the assemblage regression because we are assembling inflation subcomponents and the resulting product that we get is called albacore for adaptive learning based core inflation and so this is um my slide about the methodology I will keep it short but just to show you what we doing we have two versions so um first of all we’re using this assemblage regression which is a general generalized non- negative Rich regression to really assemble inflation subcomponents so so we call it supervised waiting because you’re really taking the inflation subcomponents as regressors and we regress it on future headline inflation and we optimize this loss function here which is basically a linear regression with a regularization term so that we can also use a really a large set of regressors and inflation subcomponents important here is uh that we want to have non- negative coefficients because we want to interpret them as weights right and they should sum to one and the second version is um a supervised trimming version so here what we do is we take the inflation sub components we order them from lowest to highest values and then we’re trying to learn which part of this distribution is important when it comes to predicting inflation um okay so we do this for the us and for the EUR area today I will focus on the Euro area but I also have two slides for the us so that we can compare maybe also um especially important the recent search and recent months we do this for different Horizons today I will focus on the 12 steps ahead because it’s a rather short to medium term that is highly relevant for monetary policy and we are using different levels of desegregation and you see that we can handle a rather low number of um inflation subcomponents obviously but we can also go very large so for the US for example in level six we have 215 components that we are assembling to get this kind of Max exim predictive core inflation measure our benchmarks are um in this forecasting exercise headline inflation and a wide range of other uh core inflation measures that are out there and we tested before covid and also after covid so we get the search what we find for the forecasting performance is that both measures um work well so in generally we find this for all levels of desegregation for um all Horizons that we test and and um we also find that mostly Alba corank is doing better and we find larger gains uh for the search and the recent periods and um I’m afraid for now you just have to trust me that this is true because I would um rather spend the rest of my time showing you the measures and the features and not the the large forecasting table but it is in the paper and in the appendix um but let’s say uh we have a good forecasting performance and we want to see where this is coming from right and so what I show you here is how our regression assembles um the subc components and so here we have level four which means we are assembling 90 um subcomponents in the Euro area and then we I sum the weight so that you can actually see something um so here in the bars you see the weights that we get from our model and in the red dots you see what is the official headline inflation weight um so two main messages first um remember that we are not excluding food and energy a priority and still we find that they get low weight or at least food half the weight that it would get because uh they are too volatile right when we take a 12 steps ahead Horizon they don’t get a lot of weight because they are highly volatile uh instead we get higher weight for services and also here for housing which is Housing Services like repairs and services for um yeah mainly repairs and rents also we also get communication and other services and we explained this by the fact that those are the components that are rather persistent or at least more persistent and so they give they’re a better indicator for medium-term developments in this case um when we look at our trimming version we find first of all that similar to a to a um usual trim mean inflation the tails are trimmed but also uh that this trimming is highly asymmetric so instead of for example trimming the lower 15% and the upper 15% we see here that our measure actually puts quite a lot of weight on the middle to upper part of this distribution okay um and we explained this or we like in the literature some arguments that could explain this are actually like um intuitively price setting behavior of firms right that um firms tend to change prices more frequently when we are increasing than when prices are decreasing um then there is this rockets and feds phenomena I think it’s from the energy and fuel price markets uh which which shows that um or which says that rocket prices tend to rise like rockets but fall like feathers um and then a more statistical um argument would be that also the monthly price distribution is in fact asymmetric so if you focus really on asymmetric trimming or only on the medium um you might get uh you might depart from from the underlying Trend especially especially also during pandemic when this distribution shifted um to to the right um so our measure looks like this I again show you hip core inflation trim mean but also now you see Alba cor ranks which is our trimming version and Alba cor comps and I think the most important um takeaway here is that due to this focus on the middle to upper part we see that we are not getting this dip after covid instead our measure shows us um some upward pressures quite early in the mid uh 2020 and then Alba cor coms for example is more in line with traditional measures but we get the Turning Point a little bit earlier so we see this disinflation and in the recent periods U which I would like to highlight here we see that we get this uh slight easing of of this fast disinflation um process that we that we had and in this slide um I focus on the different contributions of the main aggregate so you see the contribution from energy food um Nike which is non energy industrial goods and services and uh what we find is uh what was important for the search is in fact that uh we don’t exclude food and energy a priori because I mean we assign low weights to them right but still if those Dynamics are so strong that which we saw we still get positive contribution and this helps us a bit in in in getting this um this pressures bit earlier and then for the disinflation what we know is right Goods price pressures eased and also commodity price pressures East so we get this when we are at the top um in recent periods especially when you look at Alba cor coms we see this persistent pressure that we still get from services so we also find this in our measure um yeah and I think I have one more minute to show you uh the us because I think it’s two more minutes good to show show you the us because I think it’s interesting I mean the main story is the same but of course um the inflationary pressures and also the deflation is a bit uh um a disinflation sorry it’s not a deflation uh is a bit different um from what we saw in the Euro area so um I mean the main message again uh Remains the Same so we have this um albor ranks which is not showing this dip after covid and giving us uh quite some pressures in in the early or mid 2020s and again this is due to this asymmetric uh trimming that we are that we are seeing here and then for Crums again energy is basically out and food Goods get half the weight uh and more important are services and you see again more Harmony uh for for uh the periods right after covid but then also again for the disinflation we get um this early sign of the disinflation but also that it kind of it tends to be um yeah less um less fast and less strong than we saw in the EUR era right so when I show you again this decomposition you see why this is the case so for the for the us we really see that the goods price pressures were stronger and then the key drivers you see this especially for for ab corang which also helps us to get um to get this timely indication um and now recently we still see pressures from services and goods inflation so rather the core um of of our components and interesting for the us because I think uh we saw this in the second slide already that the trimp m is really lagging behind a lot and here you see why this is the case because it puts quite some weight on shelter and we know in the US shelter is kind of a uh tricky component in this inflation set because it’s a lagging indicator and you see here indeed that we get this very legging Behavior Uh when we focus so much on shelter and our measures put really um low weight on on those rather um troubling um um components so yeah um I’m already at the yend so um um I would like to conclude uh with uh the fact yeah we device two rather simple um forward-looking indicators based on this rather simple machine learning approach this is part of a broader research agenda where we always trying to redefine macroaggregate such that they satisfy certain criteria and in this case it was predictability for inflation but you can think of many other applications and about a lot a lot of um different um questions that you can answer right and we’re having an ongoing project um regarding this um being maximally predictive when it comes to what can we actually influence on the central bank and the final destination kind of would be um this assemblage Vector AO regression because then you can also put this in a rather structural setting you can have multiple macroeconomic Aggregates that you’re trying to assemble and this would be then the multivariate um and dynamic system of of this work so thanks a lot and I’m looking forward to your question we we again have time for a few questions so is there maybe yeah sure sure go ahead I think this is fascinating work um so I I wanted to understand a little bit better what you meant by out of Sample When you compute the weights and in particular um in the graphs that you were showing us both for Europe and the US did you compute the weights using data up to what quarter yeah um so thanks for pointing this out I mean I was a bit quick on on the setup this this is indeed true um so what we did in the paper is we have a certain set um where we train the data on right and then we have this out of sample period so that we can test our approach and also evaluate um forecasting performance and what I showed you here is actually estimated up to covid and then we take the weights and we apply them to in this case the last month was April 20124 and so um it’s not estimated until the end because first of all uh it’s a linear regression right you probably probably we will have to work with uh dummies or something because or with a nonlinear technique but in fact what you’re doing is I mean you always would have to end one month before right because you would need to to have this um 12 steps ahead few or whatever Horizon you choose then you compute the weights and then you can use them to get the measure so in the sense maybe I should also highlight that this is not a pure forecasting paper it’s really like we use this predi Ive um view to get to this core inflation rate so it’s really about getting the weights and then you can apply the weights um until the end of the sample or whenever you have the data thank you okay do we have more questions uh thank you so my question is a bit of a technical nature um so I was wondering you showed us very quickly how you um calculate your weights and also the loss functions you use in order to do so so my question would be if you and your co-authors considered using more of a density measure instead of this point forecast because this maybe also be interesting especially in times when we see this you know risk in inflation so maybe considering density forecast instead of Point forecast when calculating your loss might be an option or I don’t know if you considered it or looked into it yeah um thanks for the question uh so indeed here we really focus on point forecasting this is true I mean the loss function shows this we have an A quantile extension in fact so it’s not a density then it’s not a density loss function but it’s a quantile regression so that we can also focus on the lower and the upper end so this is what we did because I think it was also um at the beginning was also to to keep it rather simple because we wanted it to you know be be communicated quite easily and I think this is kind of the easiest way to communicate this measure um and also because um in the end we care a lot about the point forecast but obviously we could um extend it to a density measure as well I think um that would be possible and would be interesting for now we have the quantile approach um but yeah thanks for the suggestion okay uh more question question online maybe okay so I’m I’m going to ask you the question I’m I’ve been asking for six months now so for the moment you you are training it on the final vintage of the data now and that’s when we try to to do everything we have flash we don’t really have we have all the revision especially for PC this is quite revised deeply so uh could you train also your your your setup to work better with um just earlier vage or no casting on this kind of of of analysis or do you think that the result would be very different or do you think they would be very similar how robers do you think you are to these kind of issues yeah um yeah so for the EUR area I mean we don’t have that because the hscp does not get revised um or not like um usually there are no really large revisions but for the yes yes and we tested it after your comment in fact and it changes it a bit yes because um but in fact so the main story remains we still get low weight on energy and food components we get higher weight on services but yes it changes a bit uh but um yeah so I think my answer would be that the story remains which is good which is nice it’s like an robustness check but indeed it’s true um and I think the problem that we also have is that we kind of always need to wait for the subcomponents which are published later right so when we get the flash estimates it does not really help us at that point because we would also need the subcomponents yeah okay uh more question I think we can even close the session in advance so thank you very much and thanks all the presenter for being on time and now we we have some concluding remark by the governor just a of can time in thanks again that was again a great ftic panel and it’s for meina up to make the closure of this conference everything even the best conference find an end in this or the F conference that tried to address the challenges which AIT Central Bank uh across not only the EUR area and Beyond there and uh to work out what can be done in order to trust them and uh we started out with a great keynote speech by senos of fides who I think was very convincing to say that uh the way how we currently do monetary policy assessments uh is a bit rough because we used the nominal R in order to assess where we are and he proposed a number of uh uh a number of uh shortcuts uh which allow us a better understanding a number of ideas how to uh better allow us to assess monetary policy and I have to say I was uh uh deeply impressed by how much I to say we could learn if and as we used his proposal in our where Maria it’s over there same my idea is just to warn you that we go over and have a look how much it can bring to the discussion perhaps at the council there in order to to uh increase some of the discussions thanks yes that’s I this yes so cor thank you so thanks the second part is then we had a discussion about cbdc I think it will been ongoing discussion over the next couple of years or more it’s the question what does it bring for for central banks how much should we get engaged and uh because it has bearing on as we have seen uh on our uh financing but also our capacity to deliver and so this is one of the many panels we can expect in the future then we had the dinner speech B rastra who made clear that uh uh if and as we do monetary policy to say and we come into new areas we have to be careful not to oversteep our our mandator and uh other said that we have to be open to discussion but always come back and say is this really still our mandate and do we have the uh the authority to do so the academic sessions I won’t be able ble to summarize in a few words have to say I was always fascinated by the tri with which like on this one you brought forward your proposals it was really good to see and I think this mixture of academic uh sessions where your paper present and also then uh the uh policy sessions where the policy issues are stressed I think it’s a very good healthy part because we can mutually learn from each other and improve our assessment and uh uh in my own panel uh which dealt with the issue of for uh Rising um uh perhaps fiscal financing issues of central banks I think this was also one of the many we’ll have in the future because from what I get out I have to say I was not so convinced by some of the assessments that we really have this kind of uh uh possibility to neglect the issue and first of all because it does not really exist and be if it exists we can ignore it I think this is something which will will occupy in the future and uh then my colleague from he gave her uh a very interesting and insightful speech particular how to say the second part and where he tried uh to state yes I to say we have for um our policy monetary policy we have to focus on but also we have to have to say when VI on fiscal policy all bite we should not really get into the details and I found this conclusions pretty U insightful uh uh that also the way he did it in his letter to the president saying you know there’s a fiscal issue looming there uh outlining it without saying how to address it and that is something I think which is also uh important uh for all of our countries I think fiscal issues will be the critic question over the next couple of years and perhaps decays because with the many challenging topics we have in front and our job will be on the one hand to make sure that we are not too much impact in our monetary policy if at all on the other hand to ENC encourage uh the governments to do something uh without really how to say getting involved and uh the before last panel I won’t summarize the last one I sure you you recall it uh was about the monetary policy in the PO crisis which for me I followed it on the screen was a kind of optimistic notion there yes do say AIC crisis but the end of the day we seem to have all the instrument in order to address it so I think uh it was a very uh intensive this intensive to days I think we had great presentation and we can go home enriched and hopefully encouraged to say that we are on a good way in order to support monetary policy theoretically and their policy wise and this leads me to my last part which is always the big thanks such a conference like this starts out early require a lot of work by many people there and uh uh I clearly want to thank I to say the team of the be which is here of course the panelists these ones and the other ones and of course also uh this wer Group which here uh with anesto where is he uh oh yes he’s over there in just behind me and uh uh uh for his intellectual guidance he gave the great collaboration uh with Maria and our team and of course and the supports team uh behind which uh the event management which made everything possible including how to say that we had this great to lunches we had the dinner yesterday and with all the uh preparation which was needed there uh just to end of course all what has been said and delivered here will be on our website and uh you can I to say see it there download it and a promise I make and the name of the team is we’ll make sure that the next year’s conference will be as fascinating in these ones and I hope you will be able to join us for next year so give the team who was responsible for that a great Applause thank you [Applause] thank you