Okay, good morning, everyone. So this conference is about the future of central banking. Yesterday we had sessions on the future of macroeconomics, the future of monetary policy, which of

course are topics which would be on the programme anywhere in the world where there was a conference about the future of central banking. Of course, here in Europe, it's necessary to have a third conversation, which is relating to the unique nature of the European Monetary Union, where we have a single monetary policy, a single central bank, but of course 19

member countries. And of course, this is a much broader conversation than the role of the central bank, but the central bank, ECB, the Eurosystem, of course, has to be super aware of the nature of a monetary union and the constraints associated with the architecture at a given point in time.

So I think in discussing the future of EMU, it's especially helpful to have people who have been involved in this debate, even before the Euro was launched. So I think the design of this panel is very appropriate, given the lineup of speakers. I should say also, I just went back

and I checked to see what Vitor had written on this topic before. I wouldn't claim to have searched the full Portuguese archives from the 1990s, but I did find in the 2004 ECB central

banking conference, Vitor wrote a very interesting essay on the experience of Portugal so far. And I would encourage you to actually go back and read that. It's very much consistent with the dinner speech from Jean-Claude Trichet last night. A lot of concern about imbalances,

a lot of concern about the absence of self-correcting mechanisms within the Euro area. And of course, it's everyone looking back in time, rerunning history, what we knew then, what we know now. It's really an ongoing conversation. So this morning,

let me first turn to the first presentation, which is from Barry Arkengreen. So Barry, over to you. Thank you, Philip, and good morning, ladies and gentlemen. Management, which means both the organizing team and Vitor Constantio have asked me to

to talk about new developments in the theory of monetary unions or optimum currency area theory. So this could, in principle, be a very short talk. One way of viewing the literature is

that there has been remarkably little new theorizing about the nature of monetary unions since the advent of the Euro. One could make a stronger statement that there's been relatively little new theorizing since Mundell 1961. But I think there is an important new

strand in the literature which emphasizes financial factors, cross-border financial flows within the monetary union, and financial fragility issues that were sadly neglected by the fathers

and mothers of optimum currency area theory and could have attracted more attention as well, I think, in the run-up to the Euro in 1999. And Vitor has been a significant figure, I think, in contributing to that discussion about the role of finance, the role of

imbalances within the Eurozone and those new developments. So when I think about this topic, I'm naturally drawn, I am a prisoner of my own earlier work, as many of us are. I think this article may be my most widely cited in economics, closing in on 2000 citations

last time I looked, which is testimony to the intellectual heft of my co-author Tam Bayoumi at the International Monetary Fund. The conclusion of that piece, which was published happily exactly 25 years ago in 1993,

was that proceeding with a large monetary union encompassing not only the European core or the northern European countries but also southern Europe or peripheral Europe or the Club Med countries, as they were sometimes called at the time, was problematic.

So we built directly on the theory of optimum currency areas, such as it was, it is, which looks at a number of preconditions for the existence of a smoothly functioning monetary

union, where in this paper we looked both at the nature of business cycle disturbances and how correlated or not they were across potential constituent states of the monetary union.

We looked also at some aspects of the adjustment mechanism, whether the adjustment mechanism indicated faster adjustment, reflecting greater labor mobility, for example, in some cases than others, but it was really the symmetry or asymmetry of disturbances across the constituent member states that we focused on. The basic framework could not be simpler. It was a

straightforward aggregate supply, aggregate demand setup, in which aggregate demand shocks raise output temporarily as you shift the downward sloping aggregate demand curve to the right and prices, but over time the system rotates back toward the unchanged

vertical long run aggregate supply curve where there is no long run or permanent impact on the level of output but there is an impact on the level of prices. Contrasted with

an aggregate supply shock where you shift both the short run and the long run aggregate supply curves to the right, meaning that both output rises and prices decline permanently in the long run. So we estimated these two relationships using time series on prices and output,

country by country actually the percentage changes in prices and output. We distinguish two shocks or disturbances, one that was constrained to affect output only temporarily

but prices permanently, the aggregate so-called aggregate demand shock, and a second shock that was allowed to affect both output and prices permanently, the permanent or so-called aggregate supply shock. So we estimated this bivariate vector autoregression in prices

and output with these structural restrictions. Imposed we looked at how correlated, how symmetric or asymmetric the estimated shocks were across countries and we compared the situation

in different candidate member states in Europe with that in the United States where we distinguished a dozen or so census regions where the thought experiment for Europe was how correlated were these temporary and permanent aggregate demand and aggregate supply

shocks in different countries, how correlated were they with those in Germany, in the case of the United States the center or anchor region that we mostly worked with was the middle Atlantic states. This is what we found when we had data up through 1988.

We found that the correlation of shocks with those in the anchor country or region were lower in Europe than in the United States. So in principle you want to be in the upper right-hand

corner of these boxes where you have a high correlation of demand shocks on the vertical axis and supply shocks on the horizontal and you can see there's more such clustering in the case of US regions than there is in the case of European countries.

And you can see in the top panel as well that there was a clear distinction between two groups of European countries, a European core where the correlation of shocks was quite similar in some sense to that evident among US regions and then a second group of

countries where everybody now knows who they were, Portugal, Ireland, Italy, Greece, Spain and the UK where the correlation was noticeably lower. So that was the basis for our conclusion that a subset of European countries did not obviously fail to meet this precondition

for a smoothly operating monetary union along US lines but proceeding with a large monetary union including these so-called peripheral European countries would be much more

problematic. So here's the update using data now basically from 1994 through 2014. European countries obviously in red and US regions in blue. The red dots are still

further from the upper right-hand corner of the quasi box. Europe still looks less like optimum currency area than does the United States. Judged on this particular criterion, the red dots for Europe tend to be lower and further to the left. The US data points do not

look very different from before. There are some small differences in the correlation with mid-Atlantic reflecting I think changes in the geography of manufacturing activity in the United States over the last couple of decades. But the results for the US,

and I'll show you some further results for the US in a moment, are quite similar to what we found a quarter of a century ago. Europe looks a little bit more like an optimum currency area today than it did using the pre-1989 data. So in this figure the blue dots are European

correlations with Germany in the period up through 1988 and the red dots are the correlations with Germany for the same countries using the data from 1994. The red dots are a little bit further

to the right reflecting mainly a stronger correlation of aggregate demand shocks which some might say is a logical implication of the fact that the monetary disturbances affecting these economies are now more similar. What's unexpected in this figure is that the correlation of

disturbances, the correlation of demand shocks especially but supply shocks as well has grown more symmetric with those in Germany not in the northern European countries but in

the crisis countries. It's in Greece, in Ireland, in Spain, in that group of countries where the correlation with Germany has gone up and is greatest. So that was the big surprise of

our update. We went back and checked our code four times to be sure this was not a programming error and we are very confident at this point that it's not. So something else is going on here that needs to be understood. Our interpretation of what's going on is that this correlation

reflects capital flows between northern and southern Europe on a scale that tied these economies together on a scale that did not exist prior to the euro. So to put it in simple terms,

large capital flows from Germany to the south led those economies to boom together between 2001 and 2008 and then Germany slows down after the global financial crisis and the capital flow reverses direction in the peripheral European economies slow down as well. So we

then went on to look at how these correlations were affected by introducing financial factors into the model, into the vector autoregression. So we introduced real interest rates, we introduced real credit growth and we introduced the change in real housing prices to

link back to the discussions that we had yesterday afternoon and when you do that the correlation of disturbances between Germany and the peripheral European countries goes down to much much lower levels. So we interpret that as being consistent with the idea that it is

financial factors and financial flows that are driving this new pattern. There is in addition another twist if you have a simple two bivariate vector autoregression you can shock it and look at the impulse response functions. In our earlier study the impulse response functions

were well behaved when you subject the system to an aggregate supply shock in the U.S. as I show you here but in Europe as well you get the predicted increase in output and decline in prices. The red line there shows you the impulse response and when you subject the system to an

aggregate demand shock prices and output rise in the short run and then by constraint output falls back to its initial level. Nothing in the way we estimate this model dictates which way prices ought to move and we found it quite reassuring earlier that prices

moved in the theoretically correct expected direction in response to the temporary and permanent shocks that gave us confidence that we could loosely refer to these as aggregate supply and aggregate demand shocks. And in other words there's a correspondence between

the model and the impulse responses as you see here. Now we continue to find exactly the same thing for the United States but the impulse responses for Europe look peculiar. Now

positive supply shocks raise output but also raise prices where the textbook aggregate supply demand model suggests prices ought to go down. Positive demand shocks reduce prices where the textbook aggregate supply aggregate demand model says they should raise them. So these are

well-behaved impulse responses for the United States. Here are our initially horrifying to the authors impulse responses for Europe in the period since 1994 where the permanent shock

is in red and the temporary shock by construction is in blue. So how might we understand this hypothesis? Our hypothesis is that the positive aggregate supply shock sets off a positive aggregate demand shock and the positive or negative aggregate demand shock sets off a

negative or positive aggregate supply shock. Shifting the two curves around arbitrarily as I show you in the right hand panels is sufficient to generate a pattern that looks like this. What might the underlying economics be? Our interpretation is in terms of

hysteresis and the financial cycle. By financial cycle we mean that positive supply shocks set off a financial response that also affects demand and that positive and that

system, hence the hysteresis. Consider the left hand panel here. The story would go as follows.

A positive supply shock first raises output because plausibly consistent with arguments about the benefits of monetary union, a more stable policy environment due to the euro increases aggregate supply. That boosts productivity and profitability. That in turn raises asset prices

and sets off a lending boom. The lending boom increases aggregate demand. In the case we depict here even more than aggregate supply and the result of that large boost to aggregate demand is that you get higher prices in conjunction with this increase in aggregate

supply. So we interpret that as the pre-2008 case when peripheral European countries experienced a positive supply shock, a lending boom and higher output together with higher prices, i.e. a loss of competitiveness and shock but not coming from here. All right, so

where was I? You can run the same thought experiment in reverse post-2008. Think of a negative supply shock due to impairment of the European financial system. Lower prices also mean an asset price slump and therefore less lending. Demand falls along with supply. The

demand curve shifts to the left. The result is recession and deflation and hysteresis implies that the result is permanent. So what conclusions do I draw from this exercise?

First, it's no surprise that the euro area continues to experience difficulties. It remains further than the benchmark represented by the United States from satisfying the preconditions for an optimum currency area. Shocks are still asymmetric. Adjustment to those

shocks not shown here but shown elsewhere remains difficult. No fiscal federalism, lower levels of labor mobility. Moreover, the evidence suggests that while the euro had positive efficiency effects, the associated positive supply shock unleashed large capital flows between

northern and southern Europe, inflating asset prices in the south. The resulting lending boom boosted demand in southern Europe. That created the mirage of prosperity but also led to a permanent loss of competitiveness, suggesting obviously enough the need to do something about

this capital flow problem and its effects. It's all about financial markets, in other words, as people like Vitor Constantio have repeatedly reminded us. Thank you.

Thank you Barry and thank you also for being inside the time limit which is I think a first for this event. So Lucrezia, over to you. Good morning. It's a pleasure to be here to discuss this paper because it reminds me of my youth. It was 1993 and, you know,

we were innocent at the time and that we were looking at optimal currency areas, you know. That was a long time ago. It's also a pleasure to be here for Vitor who is my favorite member of the executive board and I can say that because there is nobody else.

Okay, good to be here. So Barry has two green conclusions. So 20 years from the establishment of the European monetary union, the Eurozone is still further from the optimal currency area

than the U.S. is. So shocks are still asymmetric. And the second conclusion is that the euro area has some efficiency effects but these efficiency effects were exactly the factors that led to large capital flows from the north to the south and loss of competitiveness. And I will call that the perils of financial integration. Now, that story has been told

before so I'm not going to bother you with my version of the story. What I'm going to try to do in this discussion is to look at the data from a much more brutal perspective to try to understand what have been the big fact about, you know, the kind of synchronization business

cycles before the inception of the euro in the euro sample and then post-crisis. And actually the facts that I want to dig out are basically two. That is very important

to distinguish between pre-crisis and post-crisis. And I think that pre-crisis, by pre-crisis it means the sample of the euro from 1999 to 2008, there were two big things. Nominal convergence and then, of course, business cycle features that,

according to my analysis, were not significantly different than in the pre-crisis sample. Now, once the crisis hits and the emu face the first large shock of his existence,

actually things change and the real asymmetries actually showed up much more forcefully than before. And, you know, there are many stories that one can say about that factor. Financial frictions leading to asymmetries, fiscal policy, lack of management, crisis management tools,

and so on. But, you know, before telling the stories, let me just point to the fact. So, that is my first fact. Okay. So, this is just a measure of asymmetries. So, it's

to the European average. So, this is what I call real asymmetries. And on the right-hand side you have nominal asymmetries, which is the same measure, but for inflation. And so, this is my first fact. And you can see that, you know, it's very important to distinguish which samples

you're talking about because otherwise you're mixing quite different things. So, we had the periods of the great moderation, which is there in the middle, where there are very little asymmetries because there was very little volatility. So, low volatilities, low asymmetries.

And then big shocks, the crisis. And again, asymmetries manifested themselves. Now, it's a quite different story for nominal because one of the main driving force pre-euro and then in the euro has been nominal convergence. Inflation rates are very similar across countries. And this, you know, there is kind of a forceful international element, but there is also the

credibility of monetary policy behind those fact, which is one of the things that actually was absent in the old currency area literature. Okay. So, the role of monetary policy as a forceful, you know, inducing credibility and so on. We know that there are other factors,

but, you know, in terms of nominal convergence, it's important and is also a fact that maybe would help us to interpret some of Barry's results on which, you know, are in terms of

real output and prices, mixing these two samples. Now, the second point I would like to make is that the real correlation across country did not change in the first ten years of the euros with respect to the past. And here I also want to revisit a paper that I wrote to

celebrate the first ten years of the euros before Lehman Brothers saying fantastically, we are an ultimate currency area. So, this is the GDP per head 1970 to 1999 for the average. And these are up to 1999, the rate of growth of what I call the blue countries.

The blue countries are not the core, okay? Because in the blue country, I also put Italy. Italy has always been very much synchronized with the euro area average. So, countries that in the 1970 had a very similar income per capita level have also countries which have moved very

much in sync through the sample. And now, you know, this is I add up a few years, so this includes the euro area sample. And this is the average. And here again are the blue countries. So, very much synchronized, not much has changed. This is not econometrics, okay? This is just a

chart. Now, what about the red countries? The red countries are not just the periphery, but I put in the red country, Finland, for example, Luxembourg and so on, small open economy, which had very different starting condition in 1970. These are the red countries.

So, historically, the red countries have been always very volatile. So, this is, you know, in the sample preceding the euro and in the euro, so there is this larger volatility, boom and bust, which characterize the, you know, the real cycles in these countries. So,

now you can put this, you can do a formal exercise. You know, I have, I'm an econometrician by background, so I like to put standard errors around stuff. And then, you know, one way of putting the standard errors around this correlation is to do the following exercise. Suppose you are in 1999, okay, and you knew the euro area path of real GDP growth, okay, from 1999 to,

you know, to 2017. Could have you anticipated, you know, the behavior of GDP of each country on the basis of the past correlations, so the historical regularities and your knowledge of the

euro area, you know, kind of realization of output. So, the answer is yes, but the uncertainty

for the red countries is so large that is very difficult. I mean, one of the reasons why the answer is yes, that the, you know, that the conditional path is not significant, significantly different from what you observe. The reason is that there is so much uncertainty because of the red countries are so volatile that these differences are not significant.

And, okay, this is my counterfact that the red countries are in the circles. You have there the actual path, which is in black. That conditional path, which is in red, is like what output would have been like if I knew the future, if I had known, you know,

the average realization of GDP. If you focus on the left of that line, which is the euro area, I mean, the big, the great recession, you can see that the difference between the conditional path and the actual path is not significant. So, but there is a lot of uncertainty. Portugal,

Ireland, Spain, Luxembourg, Greece, Finland, they're all countries which historically have had very weak association with the euro area, with the euro area, while Italy is in the other camp. So, I think the story is much more complicated. These countries where countries

have started with different initial conditions, so they were converging against something absent for the optional currency area literature that, you know, these issues of convergence growth is completely out of that model. So, that's another thing. No monetary policy, no convergence. So, you know, without those factors, it's very difficult, you know, to match that theory with

the data because, you know, that there are all these things going on. Now, you can do another formal exercise is to say, now instead of putting myself at the beginning of the euro, I'm putting myself in 2008, just before the great recession, and I'm going to ask the same questions. Could I have anticipated such heterogeneity? Well, the answer unfortunately is no.

Now, the heterogeneity has been much larger than could have anticipated. And here, I think now this is in level because I think it's clear, but I mean, I don't know if these charts are visible, but let me just tell you the story. Again, you have the bands. So,

if you see that the black line gets out of those colorful bands, that means that there is something significantly large going on. And you see that there are two countries, actually there are three countries that stand out where the black line is outside.

One is Italy, which has done much worse than could have been anticipated, you know, with the pre-euro history. The other one is Germany, which has done much, much better than could have been anticipated. And the other one is Greece, but we know the story about Greece. All the other countries have done more or less in line with this, well,

except for Ireland had a very bad crisis, but now it's back on track. Okay. It has been the kind of had a deeper crisis that could have been anticipated. So, that means that first of all, that the big shocks are different than business as usual. The optimal currency area is a theory for, you know, shocks in normal times. While this is a chart that tells you that in,

you know, facing big shocks, this is where, you know, the euro area did badly in terms of heterogeneity, but it is not about the pigs, it's not about, you know, the core and the periphery, it's much more complex. I mean, the countries that actually did differently

than could have anticipated are Italy and Germany, besides of course Greece, okay, which is an outlier in these stories. So, I think this is, I mean, I think that Barry's stories about the boom and bust and financial integration, it's a very powerful story, has been, and, you know, I'm sure that it's a big part of the explanation, but there is

another story here. And this is not a story about boom and bust. It is a story about, you know, countries, you know, with very kind of long-term stagnations, which it is Italy and countries that benefit, you know, very much from, you know, from the way in which the crisis

kind of unfolds itself. Now, the fourth point he's here, how did we do with respect to the US? Now, I mean, I'm conditioned the euro area average with respect to the US. The black line is the actual GDP per capita. The blue line is what we could have expected conditionally on US

realization and pre-crazy correlations. We have done much, much worse than expected, but by, you know, by a significant magnitude. So overall, we didn't do very well, okay. And, you know, the heterogeneity is there, but the big fact is that overall we did badly. So,

so, you know, why are the big shocks different? I mean, I can tell you a lot of stories and the stories have been told also yesterday. You know, recession with financial crisis are nasties. They lead to asymmetries. This is true everywhere, not only in currency areas,

it has been true in the US. In the euro area, this is compounded by poor crisis management capacity, pro-cyclical fiscal policy, whatever, okay. So, I don't have to explain to this audience what went wrong. Let me just tell you, give you another chart.

This is just description. It's an index. I'm out of time. I have one minute. On the left-hand side, you have deficit over GDP. On the right-hand side, you have debt over GDP. For the euro area as a whole, you have three recession. The comparison of three recession is an index. So, I put equal to 100, the beginning of the recession.

And the green is the last recession. So, we, you know, as a result of the crisis, we had a huge deficit. And from the third quarter on 2009, we started a huge consolidation, which had never been seen in the history of the euro area. Look at how huge is the,

you know, of what order of management is the comparison with the other two big recession that we have experienced in the last 30 years. So, this must be part of the story. So, what does this say about optimum currency area? I mean, I'm not sure that

Orca is a very useful model of the euro area crisis or the euro area adjustment. It does not take into account the role of monetary policy credibility. I show you the nominal convergence as a sign of, you know, a powerful thing that went on. Does not deal with countries growing at different rates, the red and the blue countries. Does not deal with financial disruption, at least

not formally. Okay? And, you know, these asymmetries have been, you know, the result of this big, you know, financial 2008 shock. So, conclusions. It might be that the euro area is dysfunctional, but for reasons which were not anticipated by the Orca literature, to function, we need to strengthen the governance. I agree with Barry. It might be,

a lot might be about finance. Thank you. Thank you, Lucretia.